1
00:00:00,000 --> 00:00:02,350


2
00:00:02,350 --> 00:00:03,430
DAVID MALAN: All right.

3
00:00:03,430 --> 00:00:04,870
Well, hello, world!

4
00:00:04,870 --> 00:00:07,810
This is CS50's now
annual workshop on how

5
00:00:07,810 --> 00:00:09,940
to prepare for technical interviews.

6
00:00:09,940 --> 00:00:13,660
This is an opportunity that we hold
traditionally on campus and nowadays,

7
00:00:13,660 --> 00:00:17,710
also on Zoom, for students here at
Harvard, at Yale University, as well

8
00:00:17,710 --> 00:00:22,300
as now anyone online who are thinking
about pursuing summertime opportunities

9
00:00:22,300 --> 00:00:25,990
in the technology world or
full-time opportunities after school

10
00:00:25,990 --> 00:00:28,570
or perhaps are pivoting later
in life and are exploring

11
00:00:28,570 --> 00:00:31,240
computer science and programming
and software engineering

12
00:00:31,240 --> 00:00:34,010
and want to ultimately
move into that world.

13
00:00:34,010 --> 00:00:37,300
This workshop is specifically on how
to prepare for the kinds of interviews

14
00:00:37,300 --> 00:00:39,440
that you might encounter in industry.

15
00:00:39,440 --> 00:00:42,370
It's one thing to take
classes like CS50 or anything

16
00:00:42,370 --> 00:00:46,438
else that sort of prepare you
indirectly for a foundation in software

17
00:00:46,438 --> 00:00:47,480
engineering and the like.

18
00:00:47,480 --> 00:00:49,522
But it's another thing to
actually ready yourself

19
00:00:49,522 --> 00:00:52,390
for the concrete opportunities
that might be of interest to you

20
00:00:52,390 --> 00:00:55,060
at a tech company or really any
other company at which you might

21
00:00:55,060 --> 00:00:56,860
be applying your technical skills.

22
00:00:56,860 --> 00:01:00,070
So thanks to CS50's
own Tommy MacWilliam,

23
00:01:00,070 --> 00:01:02,550
who was one of CS50's former
head teaching fellows today,

24
00:01:02,550 --> 00:01:04,300
we should have an
amazing workshop for you

25
00:01:04,300 --> 00:01:08,260
that should give you the mindset of how
to go about preparing for interviews,

26
00:01:08,260 --> 00:01:11,350
how to go about preparing your
resume, and really, ultimately, how

27
00:01:11,350 --> 00:01:14,770
about applying what you might learn
fairly theoretically in school

28
00:01:14,770 --> 00:01:18,640
to the actual practical world of
interviewing and ultimately working,

29
00:01:18,640 --> 00:01:19,220
as well.

30
00:01:19,220 --> 00:01:20,928
And wonderfully we're
joined here by some

31
00:01:20,928 --> 00:01:24,400
of our other former and current head
teaching fellows, Maria and Connor.

32
00:01:24,400 --> 00:01:28,870
So allow me now to introduce our friend
and former colleague, Tommy MacWilliam,

33
00:01:28,870 --> 00:01:31,600
who most recently worked
at companies like Quora.com

34
00:01:31,600 --> 00:01:34,690
and most recently his
own startup, Serenade.AI.

35
00:01:34,690 --> 00:01:37,870
Tommy is hands-down one of the
smartest people and best engineers

36
00:01:37,870 --> 00:01:38,980
I've ever met.

37
00:01:38,980 --> 00:01:42,280
Indeed, in preparing
for today's interviews,

38
00:01:42,280 --> 00:01:45,310
I got to thinking that I'm not
sure if Tommy interviewed me,

39
00:01:45,310 --> 00:01:47,000
I would get the job.

40
00:01:47,000 --> 00:01:50,560
And so it's always quite interesting
to hear about the advice he has,

41
00:01:50,560 --> 00:01:53,140
the types of experiences he's
had, because not only was

42
00:01:53,140 --> 00:01:55,420
he a former student and
a former teaching fellow,

43
00:01:55,420 --> 00:01:59,020
and now go out there in the real
world, hiring and building software,

44
00:01:59,020 --> 00:02:00,470
he's really seen it all.

45
00:02:00,470 --> 00:02:02,740
And so I think we're fortunate
to have, not only him,

46
00:02:02,740 --> 00:02:05,590
but also Maria and Connor
joining us here today.

47
00:02:05,590 --> 00:02:09,205
And so without further ado,
Tommy, welcome back to CS50.

48
00:02:09,205 --> 00:02:11,830
TOMMY MACWILLIAMS: Thanks, David,
it's always great to be back.

49
00:02:11,830 --> 00:02:14,247
I love doing this interview
workshop and helping everybody

50
00:02:14,247 --> 00:02:17,620
think about how to prepare their
resume, get ready for interviews,

51
00:02:17,620 --> 00:02:22,127
do a great job interviewing, and then
think about what to do after that.

52
00:02:22,127 --> 00:02:23,710
I've been on both sides of interviews.

53
00:02:23,710 --> 00:02:27,040
I've probably done thousands of
interviews as the interviewer.

54
00:02:27,040 --> 00:02:30,400
I've also been on the other side, and
I've also bombed interviews myself,

55
00:02:30,400 --> 00:02:32,230
so I know exactly how that feels.

56
00:02:32,230 --> 00:02:35,620
And helping with this
workshop, I'm really

57
00:02:35,620 --> 00:02:39,070
hoping to help everybody just give the
best interview performance they can

58
00:02:39,070 --> 00:02:41,740
and really demonstrate the skills
that you have as an engineer.

59
00:02:41,740 --> 00:02:43,990
So joining me today
are Maria and Connor,

60
00:02:43,990 --> 00:02:46,978
so I'll turn it over to Maria
to introduce herself, as well.

61
00:02:46,978 --> 00:02:49,270
MARIA ZLATKOVA: Hi, everyone,
super awesome to be here,

62
00:02:49,270 --> 00:02:51,040
and very excited for our session today.

63
00:02:51,040 --> 00:02:55,775
I used to be [INAUDIBLE] in
CS50 years back until 2018,

64
00:02:55,775 --> 00:02:58,840
since then have worked at
Instagram as a software engineer

65
00:02:58,840 --> 00:03:03,010
and most recently as a co-founder
of my own startup called Kangaroo.

66
00:03:03,010 --> 00:03:06,580
And I'm very excited to be here today.

67
00:03:06,580 --> 00:03:09,970
As Tommy said, I've also bombed
my very fair share of interviews.

68
00:03:09,970 --> 00:03:12,880
I have some very clear
memories of the ones I bombed.

69
00:03:12,880 --> 00:03:15,940
I've also been on the other side,
done probably hundreds, definitely not

70
00:03:15,940 --> 00:03:18,488
thousands, at Tommy's level just yet.

71
00:03:18,488 --> 00:03:20,530
But the interview process
is definitely something

72
00:03:20,530 --> 00:03:24,700
that I'm eager to help everyone
master, so I'm excited to be here.

73
00:03:24,700 --> 00:03:26,810
And I'm going to hand it off to Connor.

74
00:03:26,810 --> 00:03:28,310
CONNOR LEGGETT: Right, hi, everyone.

75
00:03:28,310 --> 00:03:29,260
I am Connor.

76
00:03:29,260 --> 00:03:35,770
I am currently a head CA
for CS50 at the college,

77
00:03:35,770 --> 00:03:40,280
and I am currently a
junior at the college.

78
00:03:40,280 --> 00:03:44,230
So I am applying for internships
for next summer, 2023--

79
00:03:44,230 --> 00:03:47,290
or yeah, 2022, sorry.

80
00:03:47,290 --> 00:03:51,940
And so I have been having several
interviews a week for the past week,

81
00:03:51,940 --> 00:03:53,320
so I'm right in the middle of it.

82
00:03:53,320 --> 00:03:56,650
And just like other
people have been saying,

83
00:03:56,650 --> 00:03:59,950
I've had some of those interviews
go very, very well and some of them

84
00:03:59,950 --> 00:04:01,970
go very, very badly.

85
00:04:01,970 --> 00:04:05,170
So I'm really excited
to learn more about how

86
00:04:05,170 --> 00:04:10,090
to help those go a little bit
better than they could, possibly,

87
00:04:10,090 --> 00:04:14,037
and also to help all of you
get started in interviews.

88
00:04:14,037 --> 00:04:17,079
TOMMY MACWILLIAMS: So today, we're
going to cover a few different things,

89
00:04:17,079 --> 00:04:20,029
sort of the entire interview
process from end to end.

90
00:04:20,029 --> 00:04:22,912
So first we're going to talk about
how to choose companies and pick

91
00:04:22,912 --> 00:04:25,120
opportunities that are great
for you and where you're

92
00:04:25,120 --> 00:04:26,620
going to have a great experience.

93
00:04:26,620 --> 00:04:28,580
Next, we're going to talk about resumes.

94
00:04:28,580 --> 00:04:32,110
I'm not going to spend a ton of time on
this, but just some quick simple rules

95
00:04:32,110 --> 00:04:34,210
for putting together a great resume.

96
00:04:34,210 --> 00:04:36,760
Then we'll talk a bit
about how to prepare

97
00:04:36,760 --> 00:04:39,170
before a tech interview happens.

98
00:04:39,170 --> 00:04:43,330
Then we'll walk through during the
interview, what's expected for you,

99
00:04:43,330 --> 00:04:46,720
as well as some tips and tricks for
doing well in the technical interview

100
00:04:46,720 --> 00:04:47,512
portion itself.

101
00:04:47,512 --> 00:04:50,470
Then we're going to have a couple of
mock interviews, so one with Maria

102
00:04:50,470 --> 00:04:51,310
and one with Connor.

103
00:04:51,310 --> 00:04:53,950
We'll simulate an actual tech interview.

104
00:04:53,950 --> 00:04:56,900
We'll use real questions
that I've used in the past,

105
00:04:56,900 --> 00:05:00,040
so these just aren't
fake, random questions,

106
00:05:00,040 --> 00:05:03,380
but these are real questions that
you could see at a tech company,

107
00:05:03,380 --> 00:05:06,330
and we'll go over some techniques there.

108
00:05:06,330 --> 00:05:10,220
And then last, we'll wrap up with some
things you can do after the interview

109
00:05:10,220 --> 00:05:15,110
and looking ahead as you improve
your interviewing skills.

110
00:05:15,110 --> 00:05:19,000
So the first thing I want to
talk about is choosing companies.

111
00:05:19,000 --> 00:05:21,970
So obviously, as you're
looking around at internships

112
00:05:21,970 --> 00:05:25,960
and full-time opportunities, there are
so many companies and opportunities

113
00:05:25,960 --> 00:05:26,680
out there.

114
00:05:26,680 --> 00:05:31,780
And so what you want to do is help
narrow that down into a small set.

115
00:05:31,780 --> 00:05:34,663
Nobody wants to be doing
500 interviews a month.

116
00:05:34,663 --> 00:05:36,580
You're going to exhaust
yourself, it's tiring.

117
00:05:36,580 --> 00:05:40,120
And so the more you can narrow down
your selection up front, more efficient

118
00:05:40,120 --> 00:05:42,320
you can be in your interview process.

119
00:05:42,320 --> 00:05:45,010
So I come up with here
just four questions

120
00:05:45,010 --> 00:05:47,530
that you can ask yourself
as you're evaluating

121
00:05:47,530 --> 00:05:51,020
the space of possible companies.

122
00:05:51,020 --> 00:05:53,110
So the first thing you'll
want to ask yourself

123
00:05:53,110 --> 00:05:58,993
is, what are you looking for out of
this internship or this full-time role?

124
00:05:58,993 --> 00:06:01,660
So for some folks, especially if
you're early on in your career,

125
00:06:01,660 --> 00:06:03,952
the thing you might be looking
for out of an internship

126
00:06:03,952 --> 00:06:05,827
is you really want to
maximize for learning.

127
00:06:05,827 --> 00:06:07,660
You really want to go
somewhere where you're

128
00:06:07,660 --> 00:06:10,270
able to have a lot of
mentorship, where you'll

129
00:06:10,270 --> 00:06:14,740
be in a space that
emphasizes learning and even

130
00:06:14,740 --> 00:06:16,750
failure as a learning mechanism.

131
00:06:16,750 --> 00:06:18,730
And that might be
different than somebody

132
00:06:18,730 --> 00:06:22,900
who, say, is five or six years into
their career, and rather than learning,

133
00:06:22,900 --> 00:06:27,250
they're looking for something more of
a leadership role or to be in a place

134
00:06:27,250 --> 00:06:30,820
where they can have more impact
and take on more responsibility.

135
00:06:30,820 --> 00:06:33,850
So I'll say that everybody
here is different.

136
00:06:33,850 --> 00:06:36,130
You might have friends
who are also interviewing,

137
00:06:36,130 --> 00:06:40,605
and maybe what they're looking for is
an opportunity in some specific city,

138
00:06:40,605 --> 00:06:41,980
and they really want to be there.

139
00:06:41,980 --> 00:06:43,810
And then you say, oh, should
I should I go there too?

140
00:06:43,810 --> 00:06:45,227
That's where my friends are going.

141
00:06:45,227 --> 00:06:45,970
Not necessarily.

142
00:06:45,970 --> 00:06:47,720
Everybody is really, really different.

143
00:06:47,720 --> 00:06:49,750
And I think when I was
interviewing in college,

144
00:06:49,750 --> 00:06:52,858
something that I kind of got bogged
down in is, well, I think I want this,

145
00:06:52,858 --> 00:06:55,400
but my friends are doing all
these things that are different,

146
00:06:55,400 --> 00:06:57,490
so maybe I should do that
too, when the reality is

147
00:06:57,490 --> 00:06:59,240
that everybody is really different.

148
00:06:59,240 --> 00:07:01,870
And so just ask yourself,
what are your goals,

149
00:07:01,870 --> 00:07:04,990
and what are you looking to
get out of your next internship

150
00:07:04,990 --> 00:07:07,960
or full-time opportunity?

151
00:07:07,960 --> 00:07:10,300
The next question that
I would ask yourself

152
00:07:10,300 --> 00:07:14,050
is, would you be excited to
work with this team every day?

153
00:07:14,050 --> 00:07:16,420
You're going to be in an
internship or a full-time role,

154
00:07:16,420 --> 00:07:19,360
you're going to be working with
these people a lot on their mission

155
00:07:19,360 --> 00:07:20,890
and on their products a lot.

156
00:07:20,890 --> 00:07:24,850
And so I'd make sure that you're as
excited as you can be about that.

157
00:07:24,850 --> 00:07:28,540
During the interview process,
you'll have a lot of opportunities

158
00:07:28,540 --> 00:07:31,810
to meet with different people on
the team and hear their stories

159
00:07:31,810 --> 00:07:33,250
and ask them questions.

160
00:07:33,250 --> 00:07:37,180
And that really is an opportunity for
you to get more information about,

161
00:07:37,180 --> 00:07:39,970
is this a team that
I'd be excited to join?

162
00:07:39,970 --> 00:07:43,330
Is their working style something
that aligns with my working style,

163
00:07:43,330 --> 00:07:45,560
their goals align with my goals?

164
00:07:45,560 --> 00:07:49,150
And so it's really a good
opportunity to ask yourself that.

165
00:07:49,150 --> 00:07:51,490
I think a similar question
here is whether or not

166
00:07:51,490 --> 00:07:54,440
your values align with
the company's values.

167
00:07:54,440 --> 00:07:58,090
So in many, many, many tech companies,
you can find on their website

168
00:07:58,090 --> 00:08:01,060
or on their careers page
something with values.

169
00:08:01,060 --> 00:08:06,580
And a value is basically something
that this company values more

170
00:08:06,580 --> 00:08:08,020
than other companies.

171
00:08:08,020 --> 00:08:11,560
So for instance, a company
value might be something like--

172
00:08:11,560 --> 00:08:14,680
an early Facebook one was,
done is better than perfect.

173
00:08:14,680 --> 00:08:18,010
And what that communicated was that
we're not a bunch of perfectionists.

174
00:08:18,010 --> 00:08:22,120
If you're the type of person who wants
to get everything perfectly correct

175
00:08:22,120 --> 00:08:24,970
and be super pedantic, then
this isn't the company for you.

176
00:08:24,970 --> 00:08:28,120
Because we're the type of people who
are going to get things done just as

177
00:08:28,120 --> 00:08:30,160
much as they need to
be, maybe 80% of the way

178
00:08:30,160 --> 00:08:33,070
there, and then ship it and
iterate and get feedback.

179
00:08:33,070 --> 00:08:35,510
And that value isn't
necessarily good or bad.

180
00:08:35,510 --> 00:08:38,840
It's just something the company has
decided about their own workflow.

181
00:08:38,840 --> 00:08:40,570
And so you can look
at the company values

182
00:08:40,570 --> 00:08:42,970
and can compare that to your
own internal values and say,

183
00:08:42,970 --> 00:08:46,550
is this going to be a good fit for me?

184
00:08:46,550 --> 00:08:48,890
The last thing to ask
yourself is whether or not

185
00:08:48,890 --> 00:08:52,730
this company size matches
what you're looking for.

186
00:08:52,730 --> 00:08:55,600
I've talked to a lot of candidates
in the interview process.

187
00:08:55,600 --> 00:08:58,650
So I'll ask them, what size
company are you looking for?

188
00:08:58,650 --> 00:08:59,900
And they'll say, I don't know.

189
00:08:59,900 --> 00:09:00,942
It doesn't really matter.

190
00:09:00,942 --> 00:09:04,830
I could join a four-person company or
I could join a 4,000-person company.

191
00:09:04,830 --> 00:09:06,420
And the answer is, it does matter.

192
00:09:06,420 --> 00:09:08,602
That's maybe a bad
answer to that question,

193
00:09:08,602 --> 00:09:11,810
since your day-to-day experience at a
four-person company and your day-to-day

194
00:09:11,810 --> 00:09:15,210
experience at a 4,000-person company,
they're just going to be really,

195
00:09:15,210 --> 00:09:15,980
really different.

196
00:09:15,980 --> 00:09:19,207
No matter how much the large company
says, oh, we're just like a small team,

197
00:09:19,207 --> 00:09:21,290
or no matter how much a
small team says, but we're

198
00:09:21,290 --> 00:09:24,560
really organized and stable like a
large company, the reality is just

199
00:09:24,560 --> 00:09:28,550
there's some things that become possible
and impossible at various company

200
00:09:28,550 --> 00:09:30,050
sizes.

201
00:09:30,050 --> 00:09:33,440
So this is a chart that
I really, really like.

202
00:09:33,440 --> 00:09:36,560
I think it was originally published
by Dustin Moskovitz at Asana.

203
00:09:36,560 --> 00:09:40,922
But it breaks down startups into
these three different categories,

204
00:09:40,922 --> 00:09:42,380
based on how many people are there.

205
00:09:42,380 --> 00:09:48,890
So you can see, at a small company,
mentorship doesn't really exist.

206
00:09:48,890 --> 00:09:50,690
People help you as you
need it, but there's

207
00:09:50,690 --> 00:09:54,560
no sort of formalized
internship program.

208
00:09:54,560 --> 00:09:57,170
Your autonomy is just like,
whatever you've got to do,

209
00:09:57,170 --> 00:09:59,990
you figure out what you have to
do, and then you go and do that.

210
00:09:59,990 --> 00:10:02,450
It could involve wearing a
variety of different hats,

211
00:10:02,450 --> 00:10:04,190
being in a variety of different roles.

212
00:10:04,190 --> 00:10:05,898
And then on the other
end of the spectrum

213
00:10:05,898 --> 00:10:09,440
is a large company, where you're going
to be hired into a specific role,

214
00:10:09,440 --> 00:10:13,310
onto a specific team, where maybe
you're a backend engineer hired

215
00:10:13,310 --> 00:10:16,400
onto the infrastructure scaling
team for the machine learning

216
00:10:16,400 --> 00:10:19,550
at this company, which
is much more narrow.

217
00:10:19,550 --> 00:10:23,810
Your agency there is, you're going
to have much more clear goals set

218
00:10:23,810 --> 00:10:25,550
by the team, set by management.

219
00:10:25,550 --> 00:10:27,798
So it's very much not to
just figure out whatever

220
00:10:27,798 --> 00:10:29,090
you need to do, and then do it.

221
00:10:29,090 --> 00:10:30,590
It's going to be,
here's our team's goals

222
00:10:30,590 --> 00:10:32,900
for the quarter, here are the
metrics we're trying to hit,

223
00:10:32,900 --> 00:10:36,180
here are the projects that we think are
going to make those metrics possible.

224
00:10:36,180 --> 00:10:38,180
And mentorship can be
much more formalized.

225
00:10:38,180 --> 00:10:40,910
You might be assigned a formal
mentor on your first day,

226
00:10:40,910 --> 00:10:44,480
have a variety of different support
systems, internal documentation,

227
00:10:44,480 --> 00:10:46,110
things like that.

228
00:10:46,110 --> 00:10:48,530
And again, something I
really want to emphasize here

229
00:10:48,530 --> 00:10:51,920
is that there is
absolutely no right answer.

230
00:10:51,920 --> 00:10:54,030
Everybody is really, really different.

231
00:10:54,030 --> 00:10:55,910
And as your career
progresses, what you're

232
00:10:55,910 --> 00:10:58,165
looking for might even change over time.

233
00:10:58,165 --> 00:11:00,290
So definitely don't look
at this chart and say, oh,

234
00:11:00,290 --> 00:11:03,320
the small company's the best or
the medium company's the best.

235
00:11:03,320 --> 00:11:06,500
It's more about what's
the best for you, not

236
00:11:06,500 --> 00:11:09,920
the objective, correct answer for
what to do after you graduate college

237
00:11:09,920 --> 00:11:12,980
or what to do as the second
company you join in your career.

238
00:11:12,980 --> 00:11:16,220
You'll probably read a lot of blog
posts on the internet claiming

239
00:11:16,220 --> 00:11:19,340
there's some objective truth here,
but I promise you, there is not.

240
00:11:19,340 --> 00:11:24,850
And so it's more about reflecting
on what you're looking for yourself.

241
00:11:24,850 --> 00:11:26,370
So that's it on choosing companies.

242
00:11:26,370 --> 00:11:29,460
I think really the takeaway here
is just spend a little bit of time

243
00:11:29,460 --> 00:11:32,820
up front narrowing down your
search and also reflecting

244
00:11:32,820 --> 00:11:35,220
on what you're looking for personally.

245
00:11:35,220 --> 00:11:38,250
As we'll see later, a really
common question in interview

246
00:11:38,250 --> 00:11:41,300
is, what are you looking for, and
why would you want to join this team?

247
00:11:41,300 --> 00:11:43,410
And so the more you can
reflect on that up front,

248
00:11:43,410 --> 00:11:47,110
the better you're going to do
on those types of questions.

249
00:11:47,110 --> 00:11:50,580
So next we're going
to jump into resumes.

250
00:11:50,580 --> 00:11:55,930
And the first thing I'll say about
resumes is they are not that important.

251
00:11:55,930 --> 00:11:59,760
So a lot of people spend a
ton of time agonizing over,

252
00:11:59,760 --> 00:12:03,120
should this be a em-dash or
a semicolon in my resume?

253
00:12:03,120 --> 00:12:04,680
Which is going to get me the job?

254
00:12:04,680 --> 00:12:08,400
The answer is that it doesn't
really matter that much.

255
00:12:08,400 --> 00:12:10,890
I think on average-- we've
talked to a lot of recruiters,

256
00:12:10,890 --> 00:12:12,510
I've been a hiring manager myself.

257
00:12:12,510 --> 00:12:14,940
I'd say the average
amount of time I spend

258
00:12:14,940 --> 00:12:17,460
reading a resume is about 30 seconds.

259
00:12:17,460 --> 00:12:22,140
And so as you're creating your resume,
really, really optimize for that.

260
00:12:22,140 --> 00:12:25,530
The person reading this is not
going to spend more than a minute

261
00:12:25,530 --> 00:12:28,090
or so reading it.

262
00:12:28,090 --> 00:12:30,630
And so what I've done is
I've created these 10 resume

263
00:12:30,630 --> 00:12:33,480
rules that are all pretty simple.

264
00:12:33,480 --> 00:12:37,668
As long as you follow them, you'll
probably have a pretty good resume.

265
00:12:37,668 --> 00:12:40,710
But before I jump into my rules, I'd
love to turn it over to the audience

266
00:12:40,710 --> 00:12:44,610
and hear, what are some resume
tips that you've heard in the past?

267
00:12:44,610 --> 00:12:47,010
DAVID MALAN: Tommy,
Martin mentions one page.

268
00:12:47,010 --> 00:12:50,610
Jessica notes that resumes
mean a lot in South Africa.

269
00:12:50,610 --> 00:12:53,100
Anna proposes that it be easy to read.

270
00:12:53,100 --> 00:12:55,530
Johnny mentions keywords.

271
00:12:55,530 --> 00:12:58,050
Rocky mentions targeting for the job.

272
00:12:58,050 --> 00:13:00,150
Assard mentions concise.

273
00:13:00,150 --> 00:13:03,390
Jessica also notes to add skills in it.

274
00:13:03,390 --> 00:13:06,390
Abhishek echoes the sentiment
that it should be one page.

275
00:13:06,390 --> 00:13:08,370
Keep it plain says, Tamika.

276
00:13:08,370 --> 00:13:13,710
Moaz says one page, also nothing
too colorful; concise and clear;

277
00:13:13,710 --> 00:13:17,310
not to make it too long; skills at
top; focus on the role you played;

278
00:13:17,310 --> 00:13:20,070
tech stack on top; clear
font; has to be concise.

279
00:13:20,070 --> 00:13:22,070
I think we can keep going
and going here, Tommy.

280
00:13:22,070 --> 00:13:24,237
TOMMY MACWILLIAMS: Yeah,
lots of great advice there.

281
00:13:24,237 --> 00:13:28,410
I love a bunch of people piling on
the concise, simple, and often takes

282
00:13:28,410 --> 00:13:29,610
the form of one page.

283
00:13:29,610 --> 00:13:33,550
I think if I had to give one piece
of advice, it would be that, as well.

284
00:13:33,550 --> 00:13:36,600
So here are my 10 rules that,
as long as you follow these,

285
00:13:36,600 --> 00:13:39,160
your resume is going
to be in good shape.

286
00:13:39,160 --> 00:13:42,300
So number 1-- one page, no exceptions.

287
00:13:42,300 --> 00:13:43,320
There's a lot of folks--

288
00:13:43,320 --> 00:13:45,570
and again, this is going to
vary a little bit by role,

289
00:13:45,570 --> 00:13:48,570
so I guess here's an
exception to my no exception.

290
00:13:48,570 --> 00:13:52,530
If you're in academia, for instance,
a really long CV is much more common.

291
00:13:52,530 --> 00:13:55,200
But if you're looking
for a typical tech job,

292
00:13:55,200 --> 00:13:58,200
one page, and don't make
any exceptions there.

293
00:13:58,200 --> 00:14:01,920
The author of Python, the person
who created the Python programming

294
00:14:01,920 --> 00:14:03,895
language, he used to be at Dropbox.

295
00:14:03,895 --> 00:14:06,270
And at some point, he published
his resume and just said,

296
00:14:06,270 --> 00:14:08,560
oh, here's what mine looks
like, and his is one page.

297
00:14:08,560 --> 00:14:11,227
And so if the person who created
the Python programming language

298
00:14:11,227 --> 00:14:16,290
can put all of his achievements and
skills onto one page, so can you.

299
00:14:16,290 --> 00:14:18,390
Number 2-- I think folks
mentioned this, as well--

300
00:14:18,390 --> 00:14:20,160
is make it really easy to skim.

301
00:14:20,160 --> 00:14:23,280
As I mentioned, the person
reading this resume,

302
00:14:23,280 --> 00:14:26,920
they're not going to spend hours
poring over every little detail.

303
00:14:26,920 --> 00:14:30,930
And so if you make it really hard
to find key information like,

304
00:14:30,930 --> 00:14:35,530
what is your name, what is your email,
what roles have you had in the past,

305
00:14:35,530 --> 00:14:38,880
what are your technical skills, if
that's hard to find on your resume,

306
00:14:38,880 --> 00:14:41,820
they could just end up
giving up on your resume

307
00:14:41,820 --> 00:14:44,020
and not taking the
time to try to find it.

308
00:14:44,020 --> 00:14:46,860
So keep in mind, the person reading
this, they're trying to go fast.

309
00:14:46,860 --> 00:14:52,150
Make it easier on the reader of your
resume to find important things.

310
00:14:52,150 --> 00:14:54,442
Number 3, make your
contact info obvious.

311
00:14:54,442 --> 00:14:57,150
I've had resumes in the past where
I've said, wow, this is great.

312
00:14:57,150 --> 00:15:00,570
Some recruiter just passed it along
to me, this resume looks great--

313
00:15:00,570 --> 00:15:02,910
no way of contacting the
person on the resume.

314
00:15:02,910 --> 00:15:06,442
There is no phone number and no email,
no link to a portfolio, anything.

315
00:15:06,442 --> 00:15:07,900
So it's like, OK, now what do I do?

316
00:15:07,900 --> 00:15:10,950
And the recruiter had to go and
track them down, which was a pain.

317
00:15:10,950 --> 00:15:13,800
So the whole point of your
resume is to get in the door

318
00:15:13,800 --> 00:15:15,263
and start the interview process.

319
00:15:15,263 --> 00:15:17,430
So make sure it's really
easy for the company you're

320
00:15:17,430 --> 00:15:20,780
applying to actually reach you.

321
00:15:20,780 --> 00:15:24,380
Number 4, you want to highlight
specific accomplishments.

322
00:15:24,380 --> 00:15:29,880
The more specific and concrete
your resume can be, the better.

323
00:15:29,880 --> 00:15:35,390
So rather than having a line on
your resume like "improved company

324
00:15:35,390 --> 00:15:37,185
metrics"-- it's really vague--

325
00:15:37,185 --> 00:15:38,060
what are the metrics?

326
00:15:38,060 --> 00:15:39,080
What does improved mean?

327
00:15:39,080 --> 00:15:40,235
Improved by how much?

328
00:15:40,235 --> 00:15:42,500
As much as you can, be
specific on your resume.

329
00:15:42,500 --> 00:15:48,170
A line that said, "improved the
performance of our critical machine

330
00:15:48,170 --> 00:15:51,920
learning ranking system
by 23% by implementing

331
00:15:51,920 --> 00:15:56,240
backend optimizations like improved
caching," that was super specific.

332
00:15:56,240 --> 00:16:00,140
I gave a metric, I said what the system
was, I said what I actually did there.

333
00:16:00,140 --> 00:16:03,590
The more specific you can be, the more
that a recruiter or a hiring manager

334
00:16:03,590 --> 00:16:06,770
can understand what your actual
impact was, and that often

335
00:16:06,770 --> 00:16:08,240
reflects well on you.

336
00:16:08,240 --> 00:16:11,840
I will say that some companies will
have nondisclosure agreements, where

337
00:16:11,840 --> 00:16:14,300
you might be working on a
project that hasn't launched yet.

338
00:16:14,300 --> 00:16:18,000
Or some metrics might be confidential
and internal to the company.

339
00:16:18,000 --> 00:16:21,170
So certainly don't break any of
those nondisclosure agreements.

340
00:16:21,170 --> 00:16:25,010
But many times, metrics of the form--

341
00:16:25,010 --> 00:16:29,558
improved performance or scaled some
system to some number of users--

342
00:16:29,558 --> 00:16:31,850
those are the types of things
that are typically public

343
00:16:31,850 --> 00:16:33,475
and that you can put on a resume.

344
00:16:33,475 --> 00:16:36,350
If you have any doubts, you can also
just ask your manager or someone

345
00:16:36,350 --> 00:16:37,610
at the internship you're at.

346
00:16:37,610 --> 00:16:39,230
Just say, hey, this is my resume.

347
00:16:39,230 --> 00:16:43,900
Is it OK if I disclose
this particular metric?

348
00:16:43,900 --> 00:16:47,320
Next, definitely include
interesting personal projects.

349
00:16:47,320 --> 00:16:50,470
So especially for folks who are a
little earlier on in their career

350
00:16:50,470 --> 00:16:53,500
and might not have experience
at multiple companies yet,

351
00:16:53,500 --> 00:16:57,010
definitely include personal
projects that you've worked on.

352
00:16:57,010 --> 00:17:00,037
Maybe that could be an interesting
final project for a course.

353
00:17:00,037 --> 00:17:02,620
Maybe it could just be something
you built in your spare time.

354
00:17:02,620 --> 00:17:04,453
But it's something
that's great to highlight

355
00:17:04,453 --> 00:17:07,240
to show that you have a lot of
passion for what you're doing

356
00:17:07,240 --> 00:17:10,450
and show off what you're
capable of building.

357
00:17:10,450 --> 00:17:12,849
I will say, I'll caution
against, if you're

358
00:17:12,849 --> 00:17:15,220
able to have a project
that's a little more unique

359
00:17:15,220 --> 00:17:18,640
and something that you designed
yourself, compared to--

360
00:17:18,640 --> 00:17:22,000
many courses have a final project
that everyone builds the same thing,

361
00:17:22,000 --> 00:17:23,460
and it's kind of standard.

362
00:17:23,460 --> 00:17:25,869
I would highlight something
that you created yourself,

363
00:17:25,869 --> 00:17:29,230
rather than some kind of standard
final project, if you're able to.

364
00:17:29,230 --> 00:17:30,650
But if you're not, that's OK too.

365
00:17:30,650 --> 00:17:32,650
Something here is certainly
better than nothing.

366
00:17:32,650 --> 00:17:35,330


367
00:17:35,330 --> 00:17:38,240
Next, no charts or ratings--

368
00:17:38,240 --> 00:17:41,420
this is maybe my number
one resume pet peeve.

369
00:17:41,420 --> 00:17:47,870
A lot of people have on their resume
some chart that says, I'm a 5 out of 10

370
00:17:47,870 --> 00:17:51,620
at JavaScript, and I'm
a 7 out of 10 at CSS.

371
00:17:51,620 --> 00:17:54,080
To me, this is just a
total waste of space.

372
00:17:54,080 --> 00:17:57,710
I have no idea what your
7 out of 10 at CSS means.

373
00:17:57,710 --> 00:17:59,960
Is the inventor of CSS
a 10, or is someone

374
00:17:59,960 --> 00:18:01,652
with three years of experience a 10?

375
00:18:01,652 --> 00:18:02,610
It just makes no sense.

376
00:18:02,610 --> 00:18:07,370
A lot of people try to do this as a
way of kind of creatively expressing

377
00:18:07,370 --> 00:18:08,480
their tech stack.

378
00:18:08,480 --> 00:18:11,030
I really don't like it
at all, and it doesn't

379
00:18:11,030 --> 00:18:15,830
communicate the information you think
is communicating, more likely than not.

380
00:18:15,830 --> 00:18:19,550
Next, I would not have a specific
objective on your resume.

381
00:18:19,550 --> 00:18:22,730
Many resumes I've seen have
at the top some objective

382
00:18:22,730 --> 00:18:27,320
that says objective, colon, to
secure a spot as a software engineer

383
00:18:27,320 --> 00:18:29,783
able to have impact on this company.

384
00:18:29,783 --> 00:18:30,950
It's a total waste of space.

385
00:18:30,950 --> 00:18:34,550
We know that the objective
of you handing me a resume

386
00:18:34,550 --> 00:18:37,430
is to get a spot at the
company that I work at.

387
00:18:37,430 --> 00:18:39,147
So it's a total waste of space.

388
00:18:39,147 --> 00:18:40,730
You've only got one page to work with.

389
00:18:40,730 --> 00:18:43,158
Putting that there doesn't
really offer a lot of value.

390
00:18:43,158 --> 00:18:45,950
Again, there might be some exceptions
here in different industries,

391
00:18:45,950 --> 00:18:48,700
where maybe it's not actually clear
what role you're applying for.

392
00:18:48,700 --> 00:18:51,742
But in the case where you've gone
through some sort of online application

393
00:18:51,742 --> 00:18:54,992
that says, here's the role I'm applying
for, or you're talking to a recruiter,

394
00:18:54,992 --> 00:18:57,300
and it's really clear what
role you're applying for,

395
00:18:57,300 --> 00:19:00,950
don't waste space with an
objective on your resume.

396
00:19:00,950 --> 00:19:04,490
Number 8, use a
professional email address.

397
00:19:04,490 --> 00:19:07,175
I remember one resume I got--

398
00:19:07,175 --> 00:19:09,050
I forget what the person's
name is-- and they

399
00:19:09,050 --> 00:19:14,185
said it was their name,
123456789, 87654321,

400
00:19:14,185 --> 00:19:15,560
and that was the email they used.

401
00:19:15,560 --> 00:19:16,977
And I was like, you've got to be--

402
00:19:16,977 --> 00:19:20,180
I instantly think less of this
person, because this is their email.

403
00:19:20,180 --> 00:19:25,130
If you're still using the same
gmail address you used 20 years ago,

404
00:19:25,130 --> 00:19:26,330
just create a new one.

405
00:19:26,330 --> 00:19:28,010
First initial, last name is great.

406
00:19:28,010 --> 00:19:29,630
Last name, first initial is great.

407
00:19:29,630 --> 00:19:30,830
Your full name is great.

408
00:19:30,830 --> 00:19:33,920
It's super easy to create a
professional looking email

409
00:19:33,920 --> 00:19:35,840
using Gmail or free email service.

410
00:19:35,840 --> 00:19:36,770
Definitely do that.

411
00:19:36,770 --> 00:19:39,910
Don't have some kind of cute joke email.

412
00:19:39,910 --> 00:19:43,130
Number 9, include relevant
links if you have them.

413
00:19:43,130 --> 00:19:47,270
So if you have a LinkedIn, a GitHub
profile, a personal portfolio,

414
00:19:47,270 --> 00:19:49,960
especially if you're
applying for a design role,

415
00:19:49,960 --> 00:19:51,430
include them on your resume.

416
00:19:51,430 --> 00:19:54,190
I think sometimes a hiring
manager, before they interview you,

417
00:19:54,190 --> 00:19:58,872
they might spend a quick 30, 60
seconds skimming through your portfolio

418
00:19:58,872 --> 00:20:01,330
so you'd have something to talk
about during the interview.

419
00:20:01,330 --> 00:20:05,080
I've certainly clicked on someone's
GitHub, saw, oh, they created

420
00:20:05,080 --> 00:20:07,517
this library for progress
bars of the command line.

421
00:20:07,517 --> 00:20:08,350
That's kind of cool.

422
00:20:08,350 --> 00:20:11,200
And I'll ask them about
that during our interview.

423
00:20:11,200 --> 00:20:14,320
So that just helps your
interviewer understand a bit more

424
00:20:14,320 --> 00:20:16,660
about your background.

425
00:20:16,660 --> 00:20:20,230
And last, definitely
don't sweat the aesthetics

426
00:20:20,230 --> 00:20:22,180
and the visuals of your resume.

427
00:20:22,180 --> 00:20:24,610
And many times simpler is much better.

428
00:20:24,610 --> 00:20:26,710
Simpler is often very easy to skim.

429
00:20:26,710 --> 00:20:31,720
But don't worry about things
like colors and photos and fonts.

430
00:20:31,720 --> 00:20:34,450
You can create a really simple
resume that gets your point across

431
00:20:34,450 --> 00:20:37,480
without having to worry
about all of that.

432
00:20:37,480 --> 00:20:40,260
So next I have, a
couple example resumes,

433
00:20:40,260 --> 00:20:43,440
and I'd love to hear from folks
in the audience things that

434
00:20:43,440 --> 00:20:48,450
are good and bad about these resumes,
taking into account these 10 resume

435
00:20:48,450 --> 00:20:49,050
rules.

436
00:20:49,050 --> 00:20:50,883
And I will note that
these are fake resumes.

437
00:20:50,883 --> 00:20:52,090
These are not real people.

438
00:20:52,090 --> 00:20:54,580
So we're not roasting anyone
on the internet right now.

439
00:20:54,580 --> 00:20:56,915
So here is the first one.

440
00:20:56,915 --> 00:20:57,790
What do people think?

441
00:20:57,790 --> 00:21:00,520
What is good and bad about this?

442
00:21:00,520 --> 00:21:03,850
DAVID MALAN: Tommy, the answers
are starting to fly in here.

443
00:21:03,850 --> 00:21:05,890
The skill graph is commented on.

444
00:21:05,890 --> 00:21:08,890
Ratings are bad, too hard to
read, the skill bar, the rating

445
00:21:08,890 --> 00:21:11,800
bar is TOO MUCH AT ONCE, in all caps.

446
00:21:11,800 --> 00:21:14,920
Bad, in a word, hard
to find information,

447
00:21:14,920 --> 00:21:17,140
charts and ratings
are bad, using charts,

448
00:21:17,140 --> 00:21:23,470
two columns are good, more used space,
font, CHARTS BAD, all caps, bars,

449
00:21:23,470 --> 00:21:26,320
scale bars equal bad, hard to skim.

450
00:21:26,320 --> 00:21:29,290
Progress bar must not be
included, BAD in all caps--

451
00:21:29,290 --> 00:21:30,910
I think I'm now encouraging it.

452
00:21:30,910 --> 00:21:33,880
And there seems to indeed
be a theme developing.

453
00:21:33,880 --> 00:21:38,180
TOMMY MACWILLIAMS: Yes, I'll go
through a few things about this.

454
00:21:38,180 --> 00:21:39,910
So the first thing
that I like about this

455
00:21:39,910 --> 00:21:42,910
is that under the work
experience and education,

456
00:21:42,910 --> 00:21:44,560
it's pretty easy to skim there.

457
00:21:44,560 --> 00:21:47,770
So I can see they worked
as a data scientist.

458
00:21:47,770 --> 00:21:49,660
For some reason, in the
two work experiences,

459
00:21:49,660 --> 00:21:52,090
they changed the order of
the role and the companies.

460
00:21:52,090 --> 00:21:53,500
That's a little weird.

461
00:21:53,500 --> 00:21:56,860
But otherwise, I like that it was really
easy to see what their experience was,

462
00:21:56,860 --> 00:21:58,390
where they worked.

463
00:21:58,390 --> 00:22:03,400
The content of the resume here,
I'd say it's reasonably specific.

464
00:22:03,400 --> 00:22:06,770
I'd say, saying that they collected
and studied interpreted large data sets

465
00:22:06,770 --> 00:22:08,455
to develop new forecasting models--

466
00:22:08,455 --> 00:22:09,580
I think that's pretty good.

467
00:22:09,580 --> 00:22:11,788
I'd say the part, I'd make
it a little more specific,

468
00:22:11,788 --> 00:22:14,500
because if you can be specific
about how big the data sets were

469
00:22:14,500 --> 00:22:16,442
or exactly what you
did to analyze them, I

470
00:22:16,442 --> 00:22:18,400
think that would make
this a little bit better.

471
00:22:18,400 --> 00:22:21,520
But I just call this
reasonably specific.

472
00:22:21,520 --> 00:22:26,170
This profile at the top, this top
left, I think this is a waste of space,

473
00:22:26,170 --> 00:22:29,758
saying that they're a highly accurate
and experienced data scientist.

474
00:22:29,758 --> 00:22:32,800
And I look at here, and they have
three years of data science experience,

475
00:22:32,800 --> 00:22:34,925
so I don't need to be told
how experienced you are.

476
00:22:34,925 --> 00:22:36,490
I can look at this from the resume.

477
00:22:36,490 --> 00:22:38,560
So I don't think that's
a good use of space.

478
00:22:38,560 --> 00:22:41,330
Then yeah, I included this one on
purpose, because it has charts.

479
00:22:41,330 --> 00:22:42,730
I don't really know
what these charts mean.

480
00:22:42,730 --> 00:22:44,022
I don't know if you do, either.

481
00:22:44,022 --> 00:22:46,540
But this person looks like
they're a fantastic team player,

482
00:22:46,540 --> 00:22:48,100
but they don't have attention to detail.

483
00:22:48,100 --> 00:22:51,100
And so like they say, well, do I want
someone who doesn't have attention

484
00:22:51,100 --> 00:22:53,020
to detail being a data scientist?

485
00:22:53,020 --> 00:22:53,710
I don't know.

486
00:22:53,710 --> 00:22:57,040
I'm not really sure this communicates
with what they wanted to communicate

487
00:22:57,040 --> 00:22:58,642
if this person were a real person.

488
00:22:58,642 --> 00:23:01,350
DAVID MALAN: Can I relay a couple
of questions now from the chat?

489
00:23:01,350 --> 00:23:02,740
TOMMY MACWILLIAMS: Go for it.

490
00:23:02,740 --> 00:23:06,220
DAVID MALAN: Sergei asks
a question about the CV.

491
00:23:06,220 --> 00:23:12,670
Do we need to adapt content in key words
for ATS, application tracking systems?

492
00:23:12,670 --> 00:23:15,280
TOMMY MACWILLIAMS: Yeah,
that's a good question.

493
00:23:15,280 --> 00:23:17,170
So in many companies,
they'll have what's

494
00:23:17,170 --> 00:23:18,550
called an ATS, which
is, as you mentioned,

495
00:23:18,550 --> 00:23:19,780
an applicant tracking system.

496
00:23:19,780 --> 00:23:24,280
So it might be Lever or Greenhouse,
just some software that you apply to.

497
00:23:24,280 --> 00:23:28,030
And some of these will kind of look at
your resume and filter out for keywords

498
00:23:28,030 --> 00:23:31,690
and, say, if it doesn't have the word
JavaScript, throw it in the trash.

499
00:23:31,690 --> 00:23:32,980
So I wouldn't go overly--

500
00:23:32,980 --> 00:23:37,600
I wouldn't go crazy and try to put
every single keyword on your resume.

501
00:23:37,600 --> 00:23:42,010
But at the same time, just having
a short list of languages you know

502
00:23:42,010 --> 00:23:43,570
or technologies you've worked with--

503
00:23:43,570 --> 00:23:47,500
I think even I have this on my
resume, just listing out the languages

504
00:23:47,500 --> 00:23:48,820
you've worked with in the past.

505
00:23:48,820 --> 00:23:53,937
If you want it to have a little bit of
this or proficient with this language,

506
00:23:53,937 --> 00:23:56,770
or I'm new to this language or
something, maybe a little bit of that

507
00:23:56,770 --> 00:23:57,190
is OK.

508
00:23:57,190 --> 00:24:00,357
But in general, just having a quick
list that doesn't take up a lot of space

509
00:24:00,357 --> 00:24:01,720
is what I'd recommend.

510
00:24:01,720 --> 00:24:04,595
DAVID MALAN: And Tommy, another
question was asked, which might vary,

511
00:24:04,595 --> 00:24:05,690
I imagine, by country.

512
00:24:05,690 --> 00:24:08,923
Provolika asks, should I add
a photograph to the resume?

513
00:24:08,923 --> 00:24:11,090
TOMMY MACWILLIAMS: Yeah,
this might vary by country.

514
00:24:11,090 --> 00:24:14,890
I would say that if you're applying to
a US tech company, I wouldn't do that.

515
00:24:14,890 --> 00:24:17,110
It just falls into the same category.

516
00:24:17,110 --> 00:24:18,780
It's a waste of space.

517
00:24:18,780 --> 00:24:21,730
You can probably change that
photo and put much more detail

518
00:24:21,730 --> 00:24:24,220
about the impact you had at
an internship or a company.

519
00:24:24,220 --> 00:24:27,178
But again, this is going to vary a
bit by country, and it's hard for me

520
00:24:27,178 --> 00:24:29,050
to speak to too much outside the US.

521
00:24:29,050 --> 00:24:31,460
But within the US, I would
definitely not do that.

522
00:24:31,460 --> 00:24:33,760
DAVID MALAN: I'll continue to chime
in, Tommy, as more questions arise.

523
00:24:33,760 --> 00:24:34,885
TOMMY MACWILLIAMS: Awesome.

524
00:24:34,885 --> 00:24:40,503
So let's jump to the next resume,
things about this that are good and bad.

525
00:24:40,503 --> 00:24:42,920
DAVID MALAN: And as the comments
start to come in, Tommy--

526
00:24:42,920 --> 00:24:48,380
too long, verbose, too convoluted,
too much text, it's a dissertation,

527
00:24:48,380 --> 00:24:49,520
unstructured.

528
00:24:49,520 --> 00:24:52,310
Is that a page from a
novel, question mark?

529
00:24:52,310 --> 00:24:55,370
Hard to read, hard to skim,
looks like a wall of text.

530
00:24:55,370 --> 00:24:58,140
Your school email will
eventually expire,

531
00:24:58,140 --> 00:25:02,570
so I wouldn't use that, hard to read,
very hard to read, hard to skim,

532
00:25:02,570 --> 00:25:04,130
it's too much.

533
00:25:04,130 --> 00:25:06,200
And there seems here to be a theme.

534
00:25:06,200 --> 00:25:11,890
In fact, Jose perhaps says it
best, TLDR, too long, didn't read.

535
00:25:11,890 --> 00:25:15,140
TOMMY MACWILLIAMS: Yeah, I think that's
sort of the same reaction I have here.

536
00:25:15,140 --> 00:25:17,630
I'll start with the things
that I like about this.

537
00:25:17,630 --> 00:25:20,855
The sort of middle section here
where it says "core expertise,"

538
00:25:20,855 --> 00:25:22,070
I actually like this.

539
00:25:22,070 --> 00:25:24,250
If you were-- this is
an engineering manager,

540
00:25:24,250 --> 00:25:26,750
and so their expertise here
might be a little bit different.

541
00:25:26,750 --> 00:25:29,990
But seeing that they have
experience in vendor relations

542
00:25:29,990 --> 00:25:33,437
or global projects and
managing different time zones,

543
00:25:33,437 --> 00:25:34,520
that's pretty interesting.

544
00:25:34,520 --> 00:25:36,590
If you're an engineer a
data scientist, a designer,

545
00:25:36,590 --> 00:25:39,440
this is probably where I'd put your
languages or technologies you've

546
00:25:39,440 --> 00:25:40,100
used in the past.

547
00:25:40,100 --> 00:25:43,100
But this is actually kind of a nice
way of-- maybe the only part of this

548
00:25:43,100 --> 00:25:46,565
that skimmable, but this is a reasonable
list where I can quickly skim.

549
00:25:46,565 --> 00:25:48,860
Yeah, my main feedback on
this resume is basically

550
00:25:48,860 --> 00:25:51,860
what everyone else mentioned, is
that there's just so much text here.

551
00:25:51,860 --> 00:25:53,780
The font is really, really small.

552
00:25:53,780 --> 00:25:55,910
It's not really skimmable.

553
00:25:55,910 --> 00:25:58,678
To find the companies they worked
at is shockingly difficult.

554
00:25:58,678 --> 00:26:00,470
Even though they're in
all capital letters,

555
00:26:00,470 --> 00:26:06,292
it's pretty hard to see
where they actually worked.

556
00:26:06,292 --> 00:26:08,750
A good thing is that their name
and the contact information

557
00:26:08,750 --> 00:26:10,490
right at the top-- that's nice.

558
00:26:10,490 --> 00:26:13,100
But then again, this sort of
top section here where they're

559
00:26:13,100 --> 00:26:16,790
writing, "merging visionary leadership
with technical software engineering

560
00:26:16,790 --> 00:26:20,970
expertise," my reaction is like,
OK, don't need to have that.

561
00:26:20,970 --> 00:26:22,770
It doesn't really add any value.

562
00:26:22,770 --> 00:26:24,950
So that that section at
the top could be removed

563
00:26:24,950 --> 00:26:28,970
in favor of making this a little bit
easier to skim, not having so much text

564
00:26:28,970 --> 00:26:30,540
crunched in here.

565
00:26:30,540 --> 00:26:31,970
So I'll pause for any questions.

566
00:26:31,970 --> 00:26:32,635
Yeah.

567
00:26:32,635 --> 00:26:34,760
DAVID MALAN: Some questions
came up from Remington.

568
00:26:34,760 --> 00:26:37,070
If I'm a non-CS major,
what would be the best

569
00:26:37,070 --> 00:26:40,640
way to add courses such as
CS50 to my resume, and perhaps

570
00:26:40,640 --> 00:26:43,223
massive open online courses
outside of college more generally?

571
00:26:43,223 --> 00:26:45,265
TOMMY MACWILLIAMS: Yeah,
that's a great question.

572
00:26:45,265 --> 00:26:48,110
I put it in that same section
where languages you've used,

573
00:26:48,110 --> 00:26:50,970
technologies you've used,
courses you've taken,

574
00:26:50,970 --> 00:26:54,170
especially for undergrads or even
people making a career shift,

575
00:26:54,170 --> 00:26:56,480
where they don't have a
lot of that experience yet.

576
00:26:56,480 --> 00:26:58,820
We'd love to see courses,
especially online courses,

577
00:26:58,820 --> 00:27:00,795
to show that you've
taken the initiative,

578
00:27:00,795 --> 00:27:02,670
and you've gained some
certain set of skills.

579
00:27:02,670 --> 00:27:07,190
So I remember when I was an undergrad, I
had, basically, name, contact, and then

580
00:27:07,190 --> 00:27:09,170
languages I know, courses I've taken.

581
00:27:09,170 --> 00:27:13,340
I'd say CS50, Introduction to Computer
Science, CS61, Systems Programming.

582
00:27:13,340 --> 00:27:16,790
And since the course number might not
mean something to anybody, so just

583
00:27:16,790 --> 00:27:20,890
highlighting the courses you've
taken as an undergrad can be great.

584
00:27:20,890 --> 00:27:23,390
DAVID MALAN: And Tommy, a couple
of folks, Pragya and Achin,

585
00:27:23,390 --> 00:27:26,390
have asked about tools to
use when writing a resume.

586
00:27:26,390 --> 00:27:30,260
Should they use an online resume
maker, LawTech, Microsoft Word,

587
00:27:30,260 --> 00:27:31,280
something else?

588
00:27:31,280 --> 00:27:34,820
TOMMY MACWILLIAMS: Yeah, I would say to
use what you're most comfortable with.

589
00:27:34,820 --> 00:27:37,258
I was a pretty big
LaTeX person in college,

590
00:27:37,258 --> 00:27:38,550
and so I was a big fan of that.

591
00:27:38,550 --> 00:27:40,040
And so my resume was used in that.

592
00:27:40,040 --> 00:27:43,040
But I wouldn't recommend like, oh,
I'm going to learn some new tool just

593
00:27:43,040 --> 00:27:45,830
for the purpose of creating this
resume, since you might end up--

594
00:27:45,830 --> 00:27:48,300
it might be harder than the
tool that you already know.

595
00:27:48,300 --> 00:27:50,822
So I'd recommend just sticking
with what you know already.

596
00:27:50,822 --> 00:27:53,780
If you don't know anything yet, even
something as simple as Google Docs

597
00:27:53,780 --> 00:27:55,755
can produce a pretty
good-looking resume.

598
00:27:55,755 --> 00:27:58,130
DAVID MALAN: And a formatting
question, too-- a few folks

599
00:27:58,130 --> 00:28:03,190
have asked about two columns versus one
column in the use of color or graphics.

600
00:28:03,190 --> 00:28:04,190
Your thoughts there too?

601
00:28:04,190 --> 00:28:07,800
TOMMY MACWILLIAMS: Yeah, a little
hard to give a blanket answer there.

602
00:28:07,800 --> 00:28:11,810
I would say, write the
content of your resume first.

603
00:28:11,810 --> 00:28:14,390
The first thing, most
important thing is the content,

604
00:28:14,390 --> 00:28:17,480
where you work, highlighting
those specific achievements,

605
00:28:17,480 --> 00:28:20,190
listing out those key skills you have.

606
00:28:20,190 --> 00:28:23,570
So do that first, and then
try a few different layouts.

607
00:28:23,570 --> 00:28:26,990
I think people with a ton of
experience or a lot of content

608
00:28:26,990 --> 00:28:30,015
might land on a different layout
than somebody who is just getting

609
00:28:30,015 --> 00:28:31,890
started and doesn't have
as much of that yet.

610
00:28:31,890 --> 00:28:36,320
So I would say, try a few things,
see what looks best for the content

611
00:28:36,320 --> 00:28:37,430
that you have.

612
00:28:37,430 --> 00:28:39,430
But if your resume is
starting to look something

613
00:28:39,430 --> 00:28:41,630
like this, just this
wall of text, then I

614
00:28:41,630 --> 00:28:43,940
would take a step back and reconsider.

615
00:28:43,940 --> 00:28:46,190
DAVID MALAN: And perhaps one
final question right now,

616
00:28:46,190 --> 00:28:49,730
amalgamated from Harsh and Kyle,
who asked questions along the lines

617
00:28:49,730 --> 00:28:54,530
of, should your resume be job-specific
or applicable to any role in tech?

618
00:28:54,530 --> 00:28:57,920
And if someone has unrelated
experience, for instance,

619
00:28:57,920 --> 00:29:00,740
from chemical engineering,
should you, or how should you

620
00:29:00,740 --> 00:29:03,110
include that on a technical resume?

621
00:29:03,110 --> 00:29:06,240
TOMMY MACWILLIAMS: Yeah,
that's a really great question.

622
00:29:06,240 --> 00:29:08,000
So I've seen the spectrum here.

623
00:29:08,000 --> 00:29:12,110
When we were at Quora
and also at Serenade,

624
00:29:12,110 --> 00:29:15,980
I've seen resumes that are
literally like for Serenade, people

625
00:29:15,980 --> 00:29:20,030
who will talk about, explicitly,
Serenade, their background,

626
00:29:20,030 --> 00:29:22,880
and why it's the right fit
for exactly our company.

627
00:29:22,880 --> 00:29:23,900
I'd say you can do that.

628
00:29:23,900 --> 00:29:24,650
You don't need to.

629
00:29:24,650 --> 00:29:27,350
I think if you're applying
to 10, 20 companies,

630
00:29:27,350 --> 00:29:31,550
doing a good job on 20 different
resumes is a lot harder than doing

631
00:29:31,550 --> 00:29:33,890
a good job on one resume.

632
00:29:33,890 --> 00:29:38,050
So where I would make it specific
is for the role, not the company.

633
00:29:38,050 --> 00:29:40,080
So if you're applying
to an engineering role,

634
00:29:40,080 --> 00:29:43,800
try to put as much engineering-related
experience as possible.

635
00:29:43,800 --> 00:29:46,860
If you're applying to a design
role, and their resume just

636
00:29:46,860 --> 00:29:48,950
has a bunch of backend
engineering stuff,

637
00:29:48,950 --> 00:29:50,700
I don't know if that
makes a ton of sense.

638
00:29:50,700 --> 00:29:53,280
I'd try to put in some design
side projects you've worked on

639
00:29:53,280 --> 00:29:57,750
or courses you've taken
or projects you've built.

640
00:29:57,750 --> 00:30:01,590
And then the question around unrelated
experience, I'd say it depends.

641
00:30:01,590 --> 00:30:03,930
I think for some people,
they've had a career in,

642
00:30:03,930 --> 00:30:06,273
let's say, I think you
mentioned chemical engineering,

643
00:30:06,273 --> 00:30:09,190
and they're looking to make a career
transition into another industry.

644
00:30:09,190 --> 00:30:11,940
If that's the case, then for sure,
show the experience you have,

645
00:30:11,940 --> 00:30:15,987
show the impact you've had at companies,
and make it clear to the recruiter,

646
00:30:15,987 --> 00:30:18,570
or make it clear to whoever
you're talking with at the company

647
00:30:18,570 --> 00:30:20,460
that you are making a career pivot.

648
00:30:20,460 --> 00:30:22,738
You've worked in this
field for some time,

649
00:30:22,738 --> 00:30:24,780
and here's why you're
looking to make this pivot,

650
00:30:24,780 --> 00:30:26,490
and here's what you're excited about.

651
00:30:26,490 --> 00:30:28,500
But definitely don't
submit a blank resume,

652
00:30:28,500 --> 00:30:32,940
just because the experience isn't
exactly in that precise field.

653
00:30:32,940 --> 00:30:36,360
At the same time, I think if
you're earlier on in your career,

654
00:30:36,360 --> 00:30:40,372
and you're an undergrad, I wouldn't
just put every club you've ever joined

655
00:30:40,372 --> 00:30:42,330
and everything you've
ever done on your resume.

656
00:30:42,330 --> 00:30:45,873
I try to keep it focused to
the role you're applying to.

657
00:30:45,873 --> 00:30:47,790
So I guess my answer is
a little bit different

658
00:30:47,790 --> 00:30:50,275
if you've had a career
that's several years,

659
00:30:50,275 --> 00:30:53,400
and you're looking to make a pivot,
versus you're just starting to graduate

660
00:30:53,400 --> 00:30:55,470
college and start to break into tech.

661
00:30:55,470 --> 00:30:56,940
DAVID MALAN: We have an
infinite supply of questions,

662
00:30:56,940 --> 00:30:59,610
but why don't we let you go on,
and I'll weave them in over time.

663
00:30:59,610 --> 00:31:02,568
TOMMY MACWILLIAMS: For sure, and I'm
sure it will save time at the end,

664
00:31:02,568 --> 00:31:03,090
as well.

665
00:31:03,090 --> 00:31:04,090
So that's it on resumes.

666
00:31:04,090 --> 00:31:09,220
Like I said, resumes are really just
a tool to get your foot in the door.

667
00:31:09,220 --> 00:31:13,390
When I say they're not that important,
that's basically what I mean is that.

668
00:31:13,390 --> 00:31:16,640
I've never been in a meeting where we're
talking about a candidate and saying,

669
00:31:16,640 --> 00:31:18,598
well, they really didn't
do well in interviews,

670
00:31:18,598 --> 00:31:20,440
but their resume font
was just so perfect,

671
00:31:20,440 --> 00:31:21,732
we've got to give them the job.

672
00:31:21,732 --> 00:31:23,648
So when I say that they're
not that important,

673
00:31:23,648 --> 00:31:26,530
I mean once you're in the door,
your resume no longer matters.

674
00:31:26,530 --> 00:31:30,172
Obviously, your experience does, but
the specific typeface on your resume

675
00:31:30,172 --> 00:31:31,130
doesn't matter anymore.

676
00:31:31,130 --> 00:31:33,310
So when I say not that
important, just to be clear,

677
00:31:33,310 --> 00:31:34,360
I don't mean spend no time on it.

678
00:31:34,360 --> 00:31:36,160
But I mean, that's the
extent to which they matter

679
00:31:36,160 --> 00:31:38,118
is that they're getting
you into the interview.

680
00:31:38,118 --> 00:31:41,830
They're not going to carry you
through the whole interview process.

681
00:31:41,830 --> 00:31:45,030
So next, let's talk about what
to do before the interview.

682
00:31:45,030 --> 00:31:48,400
So the first thing I want to
talk about is what to expect.

683
00:31:48,400 --> 00:31:52,680
So with timing, it's always
better to interview early.

684
00:31:52,680 --> 00:31:55,650
So especially if you're still
in school, whether that's

685
00:31:55,650 --> 00:31:59,970
an undergrad or a grad program, try
to interview as early as you can.

686
00:31:59,970 --> 00:32:03,150
To put concrete numbers on this,
for a summer internship starting

687
00:32:03,150 --> 00:32:07,410
in May or June, people interview
as early as the previous summer.

688
00:32:07,410 --> 00:32:12,660
So I've interviewed people in August
2020 for an interview for an internship

689
00:32:12,660 --> 00:32:15,390
starting in May 2021.

690
00:32:15,390 --> 00:32:19,407
I'd say the peak season is
around October, November

691
00:32:19,407 --> 00:32:21,990
of that year for that following
summer internship or following

692
00:32:21,990 --> 00:32:23,100
full-time opportunity.

693
00:32:23,100 --> 00:32:26,710
So as much as you can, if you can
get into that wave, that's great.

694
00:32:26,710 --> 00:32:30,390
If you can't for some reason, something
comes up, or coursework is there

695
00:32:30,390 --> 00:32:34,200
and you have to interview in,
let's say, February or March

696
00:32:34,200 --> 00:32:37,020
for a role starting in June
or May, it's still possible.

697
00:32:37,020 --> 00:32:40,345
But keep in mind that some companies
might fill up their internship class,

698
00:32:40,345 --> 00:32:42,720
and it might be harder to get
a role once a lot of people

699
00:32:42,720 --> 00:32:44,620
have already accepted
their offers there.

700
00:32:44,620 --> 00:32:46,110
So if you can, interview early.

701
00:32:46,110 --> 00:32:48,780
If you can't, it's not
the end of the world.

702
00:32:48,780 --> 00:32:52,350
Also, especially if you're in school,
plan interviewing into your schedule.

703
00:32:52,350 --> 00:32:55,950
I'm not going to say try to take
some easier courses at the same time

704
00:32:55,950 --> 00:32:58,470
you're interviewing, but if
you were to do that, that

705
00:32:58,470 --> 00:33:01,710
wouldn't hurt your interview chances.

706
00:33:01,710 --> 00:33:04,170
It can be really hard to just
assume interviews and, like,

707
00:33:04,170 --> 00:33:05,730
oh-- an interview isn't just an hour.

708
00:33:05,730 --> 00:33:07,770
You're going to spend a
lot of time practicing.

709
00:33:07,770 --> 00:33:09,390
The hour is going to be
emotionally draining.

710
00:33:09,390 --> 00:33:11,110
You're going to be exhausted after that.

711
00:33:11,110 --> 00:33:14,550
So just keep in mind that
interviewing takes a lot of time,

712
00:33:14,550 --> 00:33:17,370
and don't think of it as just
a couple hour-long meetings you

713
00:33:17,370 --> 00:33:19,050
have to worry about.

714
00:33:19,050 --> 00:33:21,260
And the last thing I say
is, look for opportunities

715
00:33:21,260 --> 00:33:23,340
that are geared at your experience.

716
00:33:23,340 --> 00:33:26,540
So in particular, if you've
only taken a few CS courses,

717
00:33:26,540 --> 00:33:30,870
you're a freshman or sophomore or
maybe early on in your CS career,

718
00:33:30,870 --> 00:33:33,560
there's plenty of internship
opportunities out there for you,

719
00:33:33,560 --> 00:33:36,770
especially at larger companies
like Google, Facebook, Microsoft.

720
00:33:36,770 --> 00:33:39,650
They have programs explicitly
designed for people who

721
00:33:39,650 --> 00:33:41,400
just have a few years of experience.

722
00:33:41,400 --> 00:33:43,040
So look out for those.

723
00:33:43,040 --> 00:33:47,030
Don't apply for the senior
manager tech lead role

724
00:33:47,030 --> 00:33:48,770
if you only have a year of experience.

725
00:33:48,770 --> 00:33:51,513
Try to find the role
that's designed for you.

726
00:33:51,513 --> 00:33:53,930
You can find these just by
looking at the job description.

727
00:33:53,930 --> 00:33:56,662
They'll often say, we're looking
for three years of experience,

728
00:33:56,662 --> 00:33:58,370
we're looking for 10
years of experience,

729
00:33:58,370 --> 00:34:00,710
we're looking for zero
use of experience.

730
00:34:00,710 --> 00:34:05,480
Even if you have 2.8 years of
experience, and it says three,

731
00:34:05,480 --> 00:34:06,860
still apply for that.

732
00:34:06,860 --> 00:34:09,568
But if you have two years of
experience, and looking for 10,

733
00:34:09,568 --> 00:34:11,360
that's probably not
the right role for you.

734
00:34:11,360 --> 00:34:13,940


735
00:34:13,940 --> 00:34:16,520
So next, just a bit on the process.

736
00:34:16,520 --> 00:34:20,260
So every company is going
to have a different process,

737
00:34:20,260 --> 00:34:23,980
but these are the things
you can probably expect.

738
00:34:23,980 --> 00:34:27,909
So the first is some
online coding challenge.

739
00:34:27,909 --> 00:34:30,940
So many companies, as
a first step, they'll

740
00:34:30,940 --> 00:34:33,489
send you a link to
some site where you can

741
00:34:33,489 --> 00:34:35,230
go on and solve some coding problem.

742
00:34:35,230 --> 00:34:36,670
Maybe it's timed.

743
00:34:36,670 --> 00:34:39,190
Maybe you're allowed to run
your code before submitting it.

744
00:34:39,190 --> 00:34:41,840
I've had things that are
just like a Google Doc,

745
00:34:41,840 --> 00:34:43,480
and they're like, here just write
your code in this Google Doc,

746
00:34:43,480 --> 00:34:45,272
and whenever you're
done, just let me know.

747
00:34:45,272 --> 00:34:49,000
I see a pretty wide spectrum here,
but that's often a first step.

748
00:34:49,000 --> 00:34:52,570
And they'll be really clear with
you around what the time limits are,

749
00:34:52,570 --> 00:34:55,315
if you're allowed to use
external resources or not.

750
00:34:55,315 --> 00:34:57,190
And if they're not, just
ask whoever gave you

751
00:34:57,190 --> 00:35:00,100
the challenge what the constraints are.

752
00:35:00,100 --> 00:35:01,810
The next is a take-home interview.

753
00:35:01,810 --> 00:35:06,640
So this will usually take the form of,
here's some larger project to build.

754
00:35:06,640 --> 00:35:08,620
Maybe you're supposed to
build it in two hours,

755
00:35:08,620 --> 00:35:11,260
maybe they're supposed
to build it in two days.

756
00:35:11,260 --> 00:35:14,410
It's going to vary, but will usually
be a bit more on the practical side,

757
00:35:14,410 --> 00:35:18,730
will be, write a little command
line app that takes this as input

758
00:35:18,730 --> 00:35:22,510
and outputs this, or write a little
web application with the login

759
00:35:22,510 --> 00:35:24,800
functionality or something like that.

760
00:35:24,800 --> 00:35:27,490
So that's another interview
type you might expect.

761
00:35:27,490 --> 00:35:32,080
The technical phone screen I'd say is
probably the most common early step

762
00:35:32,080 --> 00:35:33,400
in the interview process.

763
00:35:33,400 --> 00:35:35,920
And this is what will spend the most
time on today is technical phone screen

764
00:35:35,920 --> 00:35:38,260
and on-site, where technical
phone screen is often,

765
00:35:38,260 --> 00:35:40,720
you'll meet with someone
over the phone or over video,

766
00:35:40,720 --> 00:35:43,900
or an engineer or designer or
data scientist at that company.

767
00:35:43,900 --> 00:35:46,240
And they'll still do some
technical questions together.

768
00:35:46,240 --> 00:35:48,700
We'll ask you to write some
code to solve a problem,

769
00:35:48,700 --> 00:35:51,700
and you'll share your
screen and write that code.

770
00:35:51,700 --> 00:35:53,560
And then lastly, on-sites--

771
00:35:53,560 --> 00:35:56,230
pretty similar to technical
phone screens often.

772
00:35:56,230 --> 00:35:58,690
Many on-sites are conducted
remotely, so you'll just

773
00:35:58,690 --> 00:36:02,380
do this over video, which makes it
almost the same as a technical phone

774
00:36:02,380 --> 00:36:04,720
screen, the difference
being in an on-site,

775
00:36:04,720 --> 00:36:07,390
you'll probably meet with
a wider variety of people,

776
00:36:07,390 --> 00:36:09,310
solve a wider variety of problems.

777
00:36:09,310 --> 00:36:12,310
Maybe you'll do something that's
more design-oriented, rather

778
00:36:12,310 --> 00:36:14,560
than just writing code,
and we'll go through those.

779
00:36:14,560 --> 00:36:19,870
But these are the four types or four
interview settings you can reasonably

780
00:36:19,870 --> 00:36:23,150
expect in the interview process.

781
00:36:23,150 --> 00:36:27,200
So next I want to go through
some expectations of you.

782
00:36:27,200 --> 00:36:30,303
So before I give you mine as an
interviewer, what I expect of you,

783
00:36:30,303 --> 00:36:31,720
I'll turn it over to the audience.

784
00:36:31,720 --> 00:36:37,100
What do you think your interviewer
expects you to do in an interview?

785
00:36:37,100 --> 00:36:39,260
DAVID MALAN: Be honest,
says Colleen, Tommy.

786
00:36:39,260 --> 00:36:41,790
Talk, says Jacob.

787
00:36:41,790 --> 00:36:45,330
Communicate well is a theme here too.

788
00:36:45,330 --> 00:36:48,720
Good communication skills to
behave professionally, says Aisha.

789
00:36:48,720 --> 00:36:53,010
Be precise, clear communication,
being honest, another theme.

790
00:36:53,010 --> 00:36:56,430
Don't say "um" too much, specifically.

791
00:36:56,430 --> 00:36:59,880
Ask questions to clarify, to
know enough about the company.

792
00:36:59,880 --> 00:37:02,730
Dress up professionally,
being smart, smile,

793
00:37:02,730 --> 00:37:07,530
ask questions, research the company,
explain your own projects well.

794
00:37:07,530 --> 00:37:09,847
Good info about the
job applying, because--

795
00:37:09,847 --> 00:37:12,180
TOMMY MACWILLIAMS: Yeah, those
are all really good ones.

796
00:37:12,180 --> 00:37:14,670
I think in particular,
around communication

797
00:37:14,670 --> 00:37:19,350
is maybe the most important thing,
being able to communicate clearly

798
00:37:19,350 --> 00:37:20,770
what you're thinking.

799
00:37:20,770 --> 00:37:24,150
How you're approaching the
problem is super, super important.

800
00:37:24,150 --> 00:37:26,650
So I have a few more specific ones here.

801
00:37:26,650 --> 00:37:28,650
So the first thing I'll
be expecting you to do--

802
00:37:28,650 --> 00:37:32,080
and this is related to communication--
is be able to think through a problem.

803
00:37:32,080 --> 00:37:34,650
So the whole point of an interview
is that I'm giving you something

804
00:37:34,650 --> 00:37:36,040
that you've never seen before.

805
00:37:36,040 --> 00:37:38,340
That's by design,
because what I'm looking

806
00:37:38,340 --> 00:37:42,810
for is understanding how you would
think through this new challenge

807
00:37:42,810 --> 00:37:45,360
and how you would tackle
some new problem that

808
00:37:45,360 --> 00:37:47,200
came up if we were working together.

809
00:37:47,200 --> 00:37:49,408
So that's the first thing
I'd expect, is your ability

810
00:37:49,408 --> 00:37:51,572
to think critically and
think through a problem.

811
00:37:51,572 --> 00:37:54,030
Second, maybe this is obvious
and maybe why nobody said it,

812
00:37:54,030 --> 00:37:55,030
because it's so obvious.

813
00:37:55,030 --> 00:37:57,480
But be able to write code that works.

814
00:37:57,480 --> 00:38:01,872
Sometimes that code might not have
to be exactly correct, depending

815
00:38:01,872 --> 00:38:02,580
on the interview.

816
00:38:02,580 --> 00:38:05,830
Maybe some interviewers are just looking
for pseudocode, rather than perfectly

817
00:38:05,830 --> 00:38:06,570
functional code.

818
00:38:06,570 --> 00:38:09,112
Your interviewer will tell you
what they're looking for here,

819
00:38:09,112 --> 00:38:11,770
but we're looking for you to be
able to write code that works.

820
00:38:11,770 --> 00:38:15,707
I think if I ask a question
that you can only describe like,

821
00:38:15,707 --> 00:38:18,540
oh, here's how I would do it, and
I said, OK, let's write some code,

822
00:38:18,540 --> 00:38:21,180
and you say, I don't know
how-- that's not great.

823
00:38:21,180 --> 00:38:23,263
It happens more often
than you might expect.

824
00:38:23,263 --> 00:38:25,680
But being able to actually
write the code for the solution

825
00:38:25,680 --> 00:38:27,620
is something I'd expect.

826
00:38:27,620 --> 00:38:30,180
Next thing is I expect you
to be able to fix issues.

827
00:38:30,180 --> 00:38:33,450
So oftentimes when you're
writing code, there are bugs.

828
00:38:33,450 --> 00:38:35,670
Being able to fix those
bugs, and even better

829
00:38:35,670 --> 00:38:40,067
is identifying those bugs yourself,
but not relying on the interviewer

830
00:38:40,067 --> 00:38:42,900
to say, OK, here's your code, here
are all the things wrong with it,

831
00:38:42,900 --> 00:38:44,667
and here's how it fix
them, instead, being

832
00:38:44,667 --> 00:38:46,500
able to identify problems
with your own code

833
00:38:46,500 --> 00:38:49,920
and fixing yourself
is something I expect.

834
00:38:49,920 --> 00:38:52,060
Reasoning about runtime
is really important.

835
00:38:52,060 --> 00:38:56,670
So if I say, OK, this code looks
like it works, how fast is it?

836
00:38:56,670 --> 00:38:58,845
Being able to answer that
question is important.

837
00:38:58,845 --> 00:39:02,470
You don't have to do this sort of
like-- oh, there's always fancy math

838
00:39:02,470 --> 00:39:04,605
that you might want to
do or write a proof.

839
00:39:04,605 --> 00:39:07,140
I'm really just looking
for, is this linear,

840
00:39:07,140 --> 00:39:10,977
is this polynomial, not anything
much more fancy than that.

841
00:39:10,977 --> 00:39:12,810
And the last one, which
everybody mentioned,

842
00:39:12,810 --> 00:39:14,520
is communicate really clearly.

843
00:39:14,520 --> 00:39:18,158
So we really expect you to, whether
or not you're right or wrong,

844
00:39:18,158 --> 00:39:19,950
be able to communicate
what you're thinking

845
00:39:19,950 --> 00:39:22,950
and why you're thinking that way and
make sure that I as the interviewer

846
00:39:22,950 --> 00:39:25,520
can understand you.

847
00:39:25,520 --> 00:39:31,200
So next, just some types of
interviews that you might run into.

848
00:39:31,200 --> 00:39:34,850
So the first one is the
algorithms interview.

849
00:39:34,850 --> 00:39:39,470
So this is an interview where
the solution to the problem

850
00:39:39,470 --> 00:39:42,500
might only be like five
or 10 lines of code.

851
00:39:42,500 --> 00:39:45,530
But the challenge is
figuring out the algorithm

852
00:39:45,530 --> 00:39:48,120
and behind those 10 lines of code.

853
00:39:48,120 --> 00:39:49,970
So oftentimes in the
interview, you might

854
00:39:49,970 --> 00:39:54,440
spend 20 or 30 minutes talking
through cases together, whiteboarding,

855
00:39:54,440 --> 00:39:55,280
and thinking.

856
00:39:55,280 --> 00:39:57,642
And then once you see the
algorithm, the process

857
00:39:57,642 --> 00:40:00,350
of translating that algorithm into
code is usually pretty simple,

858
00:40:00,350 --> 00:40:02,570
and you can just do-- and sometimes the
interviewer might be like, OK, great.

859
00:40:02,570 --> 00:40:05,820
Let's just go through this really quick,
because the interesting part is over,

860
00:40:05,820 --> 00:40:08,750
because you figured out the algorithm.

861
00:40:08,750 --> 00:40:13,460
The opposite of this is a coding
interview, where there isn't really

862
00:40:13,460 --> 00:40:15,410
any fancy algorithm or any fancy trick.

863
00:40:15,410 --> 00:40:18,650
It's just like, hey, we're going to
write some code to solve a problem.

864
00:40:18,650 --> 00:40:20,930
That code needs to be really clean.

865
00:40:20,930 --> 00:40:24,170
You might be writing
50 to 100 lines of code

866
00:40:24,170 --> 00:40:27,170
and factoring things out into
functions, where the code you're writing

867
00:40:27,170 --> 00:40:30,410
might be straightforward, but there
are some interesting design challenges,

868
00:40:30,410 --> 00:40:32,490
or how you'll structure the code.

869
00:40:32,490 --> 00:40:35,240
And I will say, with these two,
your interviewer will probably not

870
00:40:35,240 --> 00:40:38,855
tell you up front, like, hey, you're
in the algorithms interview now.

871
00:40:38,855 --> 00:40:41,480
But many times behind-the-scenes
companies actually do do this.

872
00:40:41,480 --> 00:40:44,660
When they design an interview loop or
sort of your sequence of interviewers,

873
00:40:44,660 --> 00:40:46,070
they'll say, OK, the
first interviewer, you're

874
00:40:46,070 --> 00:40:48,230
going to ask an algorithms
question, second interviewer,

875
00:40:48,230 --> 00:40:49,910
you're going to ask a coding
question, third interviewer

876
00:40:49,910 --> 00:40:51,420
are you going to ask something else.

877
00:40:51,420 --> 00:40:54,710
So you probably won't
be explicitly told this,

878
00:40:54,710 --> 00:40:57,470
but you can probably tell
within the first few minutes

879
00:40:57,470 --> 00:40:59,258
which kind of interview type you're in.

880
00:40:59,258 --> 00:41:01,550
But the more important thing
here is, for our purposes,

881
00:41:01,550 --> 00:41:04,520
just making you aware that these
different types of interviews exist.

882
00:41:04,520 --> 00:41:06,560
If you find yourself
writing five lines of code,

883
00:41:06,560 --> 00:41:08,540
and you think you've solved the
problem, that might be right.

884
00:41:08,540 --> 00:41:11,373
If you find yourself writing 150
lines of code to solve the problem,

885
00:41:11,373 --> 00:41:14,000
that might be right too.

886
00:41:14,000 --> 00:41:16,820
Next is a practical interview.

887
00:41:16,820 --> 00:41:19,490
So these interviews are
intentionally designed

888
00:41:19,490 --> 00:41:21,870
to be more like real world situations.

889
00:41:21,870 --> 00:41:27,118
So rather than whiteboarding
some solution to some problem

890
00:41:27,118 --> 00:41:29,910
that you probably wouldn't actually
run into on a day-to-day basis,

891
00:41:29,910 --> 00:41:34,640
these will be, write a web application
that implements a to do-list.

892
00:41:34,640 --> 00:41:37,400
Maybe you're applying to
Asana, which is a software that

893
00:41:37,400 --> 00:41:40,692
helps you make to-do lists, and they ask
you to write a to-do list application.

894
00:41:40,692 --> 00:41:43,760
It's much more practical, because
it reflects on the day-to-day work

895
00:41:43,760 --> 00:41:45,510
you might actually be doing.

896
00:41:45,510 --> 00:41:48,050
So that's another
interview type that exists.

897
00:41:48,050 --> 00:41:50,600
Another one is systems design.

898
00:41:50,600 --> 00:41:55,310
I'll say that these are
certainly more common for folks

899
00:41:55,310 --> 00:42:00,230
who are more experienced and have
designed scalable large systems before.

900
00:42:00,230 --> 00:42:03,680
But these will often take
the shape of, let's say

901
00:42:03,680 --> 00:42:06,890
that you are designing the
back end for Google Calendar.

902
00:42:06,890 --> 00:42:08,690
Tell me about how you would do that.

903
00:42:08,690 --> 00:42:12,020
And then the question ends, and
it's your job to say, OK well first,

904
00:42:12,020 --> 00:42:13,770
let's talk about the database.

905
00:42:13,770 --> 00:42:15,530
Here's the database tables I'd create.

906
00:42:15,530 --> 00:42:18,785
And now let's talk about
the tech stack I would use.

907
00:42:18,785 --> 00:42:20,660
And these are often
really, really open-ended

908
00:42:20,660 --> 00:42:23,100
on purpose, because there is
absolutely no right answer.

909
00:42:23,100 --> 00:42:26,000
The goal is to just have a
conversation with your interviewer

910
00:42:26,000 --> 00:42:29,730
and go through the process of
creating some system together.

911
00:42:29,730 --> 00:42:31,782
So I would say, worry
about this the least,

912
00:42:31,782 --> 00:42:33,740
especially if you're
earlier on in your career,

913
00:42:33,740 --> 00:42:36,573
because these are often targeted
more at folks with more experience.

914
00:42:36,573 --> 00:42:40,310
But it's not crazy that you would get
one, even earlier on in your career.

915
00:42:40,310 --> 00:42:44,610
So just be aware that it exists, but
I'd say a word about it the least.

916
00:42:44,610 --> 00:42:46,860
And then last is a culture interview.

917
00:42:46,860 --> 00:42:49,590
So this will be an interview,
at least one interview,

918
00:42:49,590 --> 00:42:51,420
where you'll write zero code.

919
00:42:51,420 --> 00:42:54,195
You'll just talk with somebody--
maybe they're a hiring manager,

920
00:42:54,195 --> 00:42:55,695
maybe they're from HR, or maybe not.

921
00:42:55,695 --> 00:42:59,160
They're just some person that you'd
be working with on a day-to-day basis.

922
00:42:59,160 --> 00:43:03,150
And they'll ask you questions about
your prior experience, projects

923
00:43:03,150 --> 00:43:06,728
you've worked on in the past,
internships you worked on in the past.

924
00:43:06,728 --> 00:43:08,520
Maybe they'll ask a
question like, when you

925
00:43:08,520 --> 00:43:11,770
were at this company, what are some
things that the company did well

926
00:43:11,770 --> 00:43:13,770
and some things you wish
the company did better?

927
00:43:13,770 --> 00:43:16,470
Or what are some of the
biggest mistakes you made?

928
00:43:16,470 --> 00:43:19,890
Or what are the biggest things that you
learned while you're at this interview?

929
00:43:19,890 --> 00:43:22,830
And so the way to prepare
for these is really just

930
00:43:22,830 --> 00:43:25,800
reflect on your own experience
before the interview.

931
00:43:25,800 --> 00:43:29,553
Be ready to talk about-- if you
worked at Google last summer,

932
00:43:29,553 --> 00:43:30,720
be ready to talk about that.

933
00:43:30,720 --> 00:43:34,020
Talk about the projects you worked
on, talk about the obstacles

934
00:43:34,020 --> 00:43:35,850
you overcame, talk about the team.

935
00:43:35,850 --> 00:43:39,010
Don't be caught off guard if someone
asks you, how was last summer?

936
00:43:39,010 --> 00:43:42,052
You're like, oh, I don't know, actually
haven't really reflected on that.

937
00:43:42,052 --> 00:43:46,320
It's really easy to just take some time
reflecting on your previous experience

938
00:43:46,320 --> 00:43:47,980
and also what you're looking for next.

939
00:43:47,980 --> 00:43:50,355
That's sort of the other half
of these questions will be,

940
00:43:50,355 --> 00:43:51,870
what do you want out of this role?

941
00:43:51,870 --> 00:43:53,110
What are you looking for?

942
00:43:53,110 --> 00:43:56,010
And remember those questions that we
started with, questions to ask yourself

943
00:43:56,010 --> 00:43:56,940
before the interview.

944
00:43:56,940 --> 00:43:58,770
If you're asking yourself
all those questions,

945
00:43:58,770 --> 00:44:01,080
you're probably going to be ready for
that part of the culture interview.

946
00:44:01,080 --> 00:44:03,705
Because you've already you've
picked this company for a reason,

947
00:44:03,705 --> 00:44:05,370
and so you're ready to answer.

948
00:44:05,370 --> 00:44:08,340
DAVID MALAN: Tommy, a good question
that came in was from Sara.

949
00:44:08,340 --> 00:44:12,360
How do act if you can't solve a
problem or can't write some code?

950
00:44:12,360 --> 00:44:13,862
This scares Sara a lot.

951
00:44:13,862 --> 00:44:15,570
TOMMY MACWILLIAMS:
It's a great question.

952
00:44:15,570 --> 00:44:16,450
I've been there.

953
00:44:16,450 --> 00:44:18,950
I think probably
everybody has been there.

954
00:44:18,950 --> 00:44:20,700
So I would say that
if you get a question,

955
00:44:20,700 --> 00:44:22,440
and you're not sure how to solve it--

956
00:44:22,440 --> 00:44:25,440
well, I'll say most people when they
first get the question are not sure

957
00:44:25,440 --> 00:44:26,140
how to solve it.

958
00:44:26,140 --> 00:44:28,620
And so we'll go through
this, but the techniques

959
00:44:28,620 --> 00:44:32,797
are really about thinking out loud,
going through your thought process.

960
00:44:32,797 --> 00:44:35,130
You might not be exactly sure,
but you can say something

961
00:44:35,130 --> 00:44:38,830
like, OK, well, we know
this list is sorted,

962
00:44:38,830 --> 00:44:41,970
so let's think about the different
algorithms for searching a list.

963
00:44:41,970 --> 00:44:43,185
There's binary search.

964
00:44:43,185 --> 00:44:44,950
Will binary search work here?

965
00:44:44,950 --> 00:44:45,820
Well, let's see.

966
00:44:45,820 --> 00:44:48,510
Just think out loud like that.

967
00:44:48,510 --> 00:44:51,190
And if you hit a wall--
which you might hit a wall--

968
00:44:51,190 --> 00:44:54,180
hopefully, your interviewer will
be proactive about guiding you

969
00:44:54,180 --> 00:44:55,350
along and giving you hints.

970
00:44:55,350 --> 00:44:57,330
But sometimes you just hit a
wall, and your interviewer is just

971
00:44:57,330 --> 00:44:59,622
kind of staring at you, and
you can just say, you know,

972
00:44:59,622 --> 00:45:02,250
I'm honestly not sure
what to think about next.

973
00:45:02,250 --> 00:45:05,020
Can you give me a hint,
or can you help me out?

974
00:45:05,020 --> 00:45:07,350
That's much better than
just spinning your wheels

975
00:45:07,350 --> 00:45:10,870
and staring at the wall in silence.

976
00:45:10,870 --> 00:45:11,370
Awesome.

977
00:45:11,370 --> 00:45:14,650
So going back into
preparing for interviews,

978
00:45:14,650 --> 00:45:17,610
so definitely don't go
into interview cold.

979
00:45:17,610 --> 00:45:20,400
Always do some practice
problems beforehand.

980
00:45:20,400 --> 00:45:23,820
The amount of preparation you put
into an interview really, really

981
00:45:23,820 --> 00:45:25,862
makes a big difference in your success.

982
00:45:25,862 --> 00:45:28,320
Some people go so far as to
say that interviewing is really

983
00:45:28,320 --> 00:45:29,790
just a learned skill.

984
00:45:29,790 --> 00:45:31,232
Nobody is good at it by design.

985
00:45:31,232 --> 00:45:33,940
Everybody just-- the more you
practice, the better at it you get.

986
00:45:33,940 --> 00:45:37,420
And it's just the skill that
you can have by practicing it.

987
00:45:37,420 --> 00:45:40,390
So we'll definitely say
that the more you practice,

988
00:45:40,390 --> 00:45:44,040
the more that is probably going to
decrease your stress in the interview,

989
00:45:44,040 --> 00:45:44,700
as well.

990
00:45:44,700 --> 00:45:47,700
Everybody knows that interviewing,
it's really stressful.

991
00:45:47,700 --> 00:45:49,470
It's scary, there's a lot of anxiety.

992
00:45:49,470 --> 00:45:52,710
The more you've practiced in advance
and the more problems you've seen,

993
00:45:52,710 --> 00:45:54,747
the more times you've
been in this situation

994
00:45:54,747 --> 00:45:56,580
with somebody asking
you an unknown question

995
00:45:56,580 --> 00:45:58,320
and having to think out loud and type--

996
00:45:58,320 --> 00:46:00,550
the more you've done
that, the more used to it

997
00:46:00,550 --> 00:46:03,090
you get, and hopefully, the
less stressful it becomes.

998
00:46:03,090 --> 00:46:05,550
So definitely do a lot
of that in advance.

999
00:46:05,550 --> 00:46:08,190
The second thing I do in
advance is pick a language

1000
00:46:08,190 --> 00:46:10,560
you're going to
interview in, and then do

1001
00:46:10,560 --> 00:46:12,700
all of your practice in that language.

1002
00:46:12,700 --> 00:46:14,460
So let's say that you like Python.

1003
00:46:14,460 --> 00:46:15,570
Great.

1004
00:46:15,570 --> 00:46:17,520
Learn Python really well.

1005
00:46:17,520 --> 00:46:20,730
You shouldn't have to Google, how
do I reverse a list in Python,

1006
00:46:20,730 --> 00:46:23,700
or how do I get the length of a
list, or how do I create a class.

1007
00:46:23,700 --> 00:46:26,550
If you have to spend any
time thinking about that,

1008
00:46:26,550 --> 00:46:28,170
it just create stress for you.

1009
00:46:28,170 --> 00:46:30,150
You just want to have
this totally down cold.

1010
00:46:30,150 --> 00:46:33,900
Don't even worry about how to create
a set in Python, because you've just

1011
00:46:33,900 --> 00:46:35,520
memorized it in advance.

1012
00:46:35,520 --> 00:46:39,578
And I would say choose one language,
maybe two, for some reason,

1013
00:46:39,578 --> 00:46:42,120
if maybe you want a backend
language and a frontend language.

1014
00:46:42,120 --> 00:46:44,040
But you should never
be in a situation where

1015
00:46:44,040 --> 00:46:47,908
an interviewer asks you a question, and
your first reaction is, what language

1016
00:46:47,908 --> 00:46:49,200
do I want to use to solve this?

1017
00:46:49,200 --> 00:46:52,275
Just have this down already
and be ready with a language.

1018
00:46:52,275 --> 00:46:55,440


1019
00:46:55,440 --> 00:47:00,920
So some things that you should know how
to do in whatever language you pick--

1020
00:47:00,920 --> 00:47:02,252
one is create a function.

1021
00:47:02,252 --> 00:47:05,210
Again, this sounds obvious, but you'd
be surprised how many people I've

1022
00:47:05,210 --> 00:47:06,950
interviewed who the
first thing they asked me

1023
00:47:06,950 --> 00:47:08,540
was, how do I create a function again?

1024
00:47:08,540 --> 00:47:10,415
Or they had to Google
how to make a function.

1025
00:47:10,415 --> 00:47:12,920
You should just know this in advance.

1026
00:47:12,920 --> 00:47:15,500
Similarly, for defining
a class, you might not

1027
00:47:15,500 --> 00:47:17,300
use a lot of
object-oriented programming,

1028
00:47:17,300 --> 00:47:20,050
but sometimes it can be helpful,
depending on what language you're

1029
00:47:20,050 --> 00:47:22,190
using, to have a class
to represent some object

1030
00:47:22,190 --> 00:47:24,090
or represent some collection of data.

1031
00:47:24,090 --> 00:47:27,050
So just know how to do that in advance.

1032
00:47:27,050 --> 00:47:29,360
A lot of interview
questions are string based.

1033
00:47:29,360 --> 00:47:33,500
And so when I say work with strings,
I mean how to create a string,

1034
00:47:33,500 --> 00:47:36,050
how to concatenate strings,
how to index into a string,

1035
00:47:36,050 --> 00:47:39,710
how to make a string uppercase or
lowercase, how to get a sub-string.

1036
00:47:39,710 --> 00:47:43,340
All of these different standard
library operations with strings,

1037
00:47:43,340 --> 00:47:44,510
just know them.

1038
00:47:44,510 --> 00:47:48,472
Again, the more you have to think
about, like, is it dot size or count,

1039
00:47:48,472 --> 00:47:50,180
it's just going to
make you more nervous.

1040
00:47:50,180 --> 00:47:51,972
You're going to create
stress for yourself.

1041
00:47:51,972 --> 00:47:53,930
And knowing all the stuff
down cold in advance

1042
00:47:53,930 --> 00:47:57,510
can just make your
interview go a lot better.

1043
00:47:57,510 --> 00:48:00,230
Same thing with lists-- many,
many interview questions have,

1044
00:48:00,230 --> 00:48:02,330
take the form, you're given a list.

1045
00:48:02,330 --> 00:48:06,170
Those are the first four words
of many interview questions are,

1046
00:48:06,170 --> 00:48:07,130
you're given a list.

1047
00:48:07,130 --> 00:48:09,920
And so just being able to be
aware of all these different list

1048
00:48:09,920 --> 00:48:12,530
operations like list
slicing or concatenation

1049
00:48:12,530 --> 00:48:14,990
or removing indexes or
pushing and popping,

1050
00:48:14,990 --> 00:48:17,990
all of those operations
you might do to a list,

1051
00:48:17,990 --> 00:48:20,750
you don't want to be asking
yourself, is it push or is it append?

1052
00:48:20,750 --> 00:48:21,410
I forget.

1053
00:48:21,410 --> 00:48:23,680
Just know that.

1054
00:48:23,680 --> 00:48:24,980
Same thing with trees--

1055
00:48:24,980 --> 00:48:26,980
another really, really
common interview question

1056
00:48:26,980 --> 00:48:30,550
is you're given a binary tree,
you're given some sort of tree,

1057
00:48:30,550 --> 00:48:32,510
just do some sort of traversal on it.

1058
00:48:32,510 --> 00:48:35,830
So being able to write depth-first
search or breadth-first search,

1059
00:48:35,830 --> 00:48:38,740
even create a tree in
whatever language you're in,

1060
00:48:38,740 --> 00:48:41,948
just knowing how to do that in advance
is going to save you a lot of trouble.

1061
00:48:41,948 --> 00:48:43,700


1062
00:48:43,700 --> 00:48:46,338
So next around complexity--

1063
00:48:46,338 --> 00:48:48,130
we mentioned that many
interviews will just

1064
00:48:48,130 --> 00:48:51,160
ask you after you've written code,
what's the complexity of this code?

1065
00:48:51,160 --> 00:48:52,765
It's pretty rare.

1066
00:48:52,765 --> 00:48:54,640
I'll say it's not
impossible, but pretty rare

1067
00:48:54,640 --> 00:48:58,720
that the complexity
they're looking for is n

1068
00:48:58,720 --> 00:49:02,770
cubed plus n factorial over something.

1069
00:49:02,770 --> 00:49:06,370
It's usually just going to be,
is it linear, is it polynomial,

1070
00:49:06,370 --> 00:49:08,380
is it exponential?

1071
00:49:08,380 --> 00:49:10,160
And that's probably it.

1072
00:49:10,160 --> 00:49:13,630
So being able to have these down
cold is really, really helpful,

1073
00:49:13,630 --> 00:49:17,320
whether something is constant, log,
linear, polynomial, or exponential.

1074
00:49:17,320 --> 00:49:21,130
Just knowing these five
and being able to identify,

1075
00:49:21,130 --> 00:49:23,200
is your code one of
these five, that's going

1076
00:49:23,200 --> 00:49:26,470
to get you through the vast
majority of what interviewers

1077
00:49:26,470 --> 00:49:31,840
are looking for when they say, what's
the complexity of your solution?

1078
00:49:31,840 --> 00:49:35,040
So next thing I'll say is
to build up your toolbox.

1079
00:49:35,040 --> 00:49:37,890
So many, many, many
interview problems are

1080
00:49:37,890 --> 00:49:43,720
kind of fancy ways of phrasing or
framing the same few core concepts.

1081
00:49:43,720 --> 00:49:44,740
So here are a few.

1082
00:49:44,740 --> 00:49:47,370
So one is recursion
or divide and conquer,

1083
00:49:47,370 --> 00:49:49,800
basically taking some
large problem, being

1084
00:49:49,800 --> 00:49:52,680
able to split it up
into smaller subproblems

1085
00:49:52,680 --> 00:49:54,570
and then solving those
smaller subproblems.

1086
00:49:54,570 --> 00:49:57,888
So as you're doing practice problems,
be really familiar with recursion.

1087
00:49:57,888 --> 00:50:00,180
And I think recursion is
something that for many people

1088
00:50:00,180 --> 00:50:02,620
is really scary, especially
if you're early on.

1089
00:50:02,620 --> 00:50:04,590
So if it is scary for
you, just practice it.

1090
00:50:04,590 --> 00:50:08,578
You don't want to be caught in a
situation where you get a problem that

1091
00:50:08,578 --> 00:50:10,620
probably should be solved
recursively, but you're

1092
00:50:10,620 --> 00:50:12,270
uncomfortable with recursion.

1093
00:50:12,270 --> 00:50:15,662
So the more you practice it, the
more comfortable you're going to be.

1094
00:50:15,662 --> 00:50:17,370
Definitely don't be
put in that situation

1095
00:50:17,370 --> 00:50:21,330
because you just didn't
want to practice recursion.

1096
00:50:21,330 --> 00:50:22,920
Next one is graph searches.

1097
00:50:22,920 --> 00:50:25,980
We mentioned this with
trees, but many problems just

1098
00:50:25,980 --> 00:50:29,940
reduced to, just model this
thing as a tree or a graph,

1099
00:50:29,940 --> 00:50:31,860
and then just do a
search over that tree.

1100
00:50:31,860 --> 00:50:34,572
So being able to practice
writing depth-first search

1101
00:50:34,572 --> 00:50:37,030
or breadth-first search, just
do it a few times in advance.

1102
00:50:37,030 --> 00:50:39,370
Make sure you're comfortable with it.

1103
00:50:39,370 --> 00:50:40,960
Greedy algorithms is another one.

1104
00:50:40,960 --> 00:50:43,627
As you're sort of looking around
at interview practice problems,

1105
00:50:43,627 --> 00:50:45,428
this is often the
category, so just being

1106
00:50:45,428 --> 00:50:48,220
able to practice these and recognize
when something might be there.

1107
00:50:48,220 --> 00:50:50,080
We talked about strings.

1108
00:50:50,080 --> 00:50:53,630
Many problems are just string
manipulations at the end of the day.

1109
00:50:53,630 --> 00:50:55,510
So really, as you're
practicing, you should

1110
00:50:55,510 --> 00:50:59,590
be building up this
toolbox as you go and just

1111
00:50:59,590 --> 00:51:01,798
be familiar with all these concepts.

1112
00:51:01,798 --> 00:51:04,090
And searching and sorting is
another really common one.

1113
00:51:04,090 --> 00:51:05,632
How do I sort this array efficiently?

1114
00:51:05,632 --> 00:51:09,110
Given assorted array, how
do I search it efficiently?

1115
00:51:09,110 --> 00:51:11,980
And dynamic programming-- this
one's a little more advanced.

1116
00:51:11,980 --> 00:51:14,480
But if you're taking algorithms
classes, dynamic programming

1117
00:51:14,480 --> 00:51:18,350
is basically a way to speed up
algorithms and sort of cache

1118
00:51:18,350 --> 00:51:22,600
the results of problems
as you're solving them.

1119
00:51:22,600 --> 00:51:26,350
So some tools you have
at your disposal--

1120
00:51:26,350 --> 00:51:28,800
so when you get a problem,
you should really think about,

1121
00:51:28,800 --> 00:51:30,000
what tools do I have?

1122
00:51:30,000 --> 00:51:32,070
What core algorithms do I know?

1123
00:51:32,070 --> 00:51:34,200
What are some core data
structures that I know?

1124
00:51:34,200 --> 00:51:38,130
And then try to pattern match which
of these data structures or algorithms

1125
00:51:38,130 --> 00:51:39,880
fits this problem.

1126
00:51:39,880 --> 00:51:41,820
So one is arrays or linked lists.

1127
00:51:41,820 --> 00:51:45,030
Be really comfortable using arrays,
understanding what a linked list is,

1128
00:51:45,030 --> 00:51:49,110
and understanding what problems should
fit into a linked list or an array.

1129
00:51:49,110 --> 00:51:50,742
Hash table is another super common one.

1130
00:51:50,742 --> 00:51:52,950
You can sort of ask yourself,
OK, given this problem,

1131
00:51:52,950 --> 00:51:55,000
can I model this as a hash table?

1132
00:51:55,000 --> 00:51:58,320
Does that, therefore,
instantly speed up my runtime?

1133
00:51:58,320 --> 00:51:59,970
Binary search is another one.

1134
00:51:59,970 --> 00:52:02,040
Many, many, many problems
that take the form

1135
00:52:02,040 --> 00:52:06,540
of find X, or given a
list, look for something,

1136
00:52:06,540 --> 00:52:08,498
can be reduced to a
binary search problem.

1137
00:52:08,498 --> 00:52:09,540
So just be aware of that.

1138
00:52:09,540 --> 00:52:13,350
And maybe that should be your first
check if you get a search problem is,

1139
00:52:13,350 --> 00:52:15,763
can I model this as binary search?

1140
00:52:15,763 --> 00:52:17,430
Shortest path algorithms or another one.

1141
00:52:17,430 --> 00:52:19,513
I think given that we
mentioned these a few times.

1142
00:52:19,513 --> 00:52:20,590
We've given some tree.

1143
00:52:20,590 --> 00:52:23,280
How do I search through
that tree efficiently?

1144
00:52:23,280 --> 00:52:26,190
And memoization, which is
similar to dynamic programming,

1145
00:52:26,190 --> 00:52:31,140
but once I've solved the problem,
can I reuse that solution

1146
00:52:31,140 --> 00:52:32,970
to solve some larger problem?

1147
00:52:32,970 --> 00:52:37,530
And we'll go into some examples of this
once you get into the mock interviews.

1148
00:52:37,530 --> 00:52:40,320
So some more general advice--
as you're practicing,

1149
00:52:40,320 --> 00:52:44,220
try to really simulate the
environment accurately.

1150
00:52:44,220 --> 00:52:47,818
So something that I know I do when
I'm practicing is I look at a problem,

1151
00:52:47,818 --> 00:52:50,110
and I'm like, oh, yeah, I
think I can basically see it.

1152
00:52:50,110 --> 00:52:53,402
And then I'll look at the answer and be
like, oh, yeah, I could have done that.

1153
00:52:53,402 --> 00:52:56,130
But then if I tried to force
myself not to look at the answer,

1154
00:52:56,130 --> 00:52:58,530
I'm like, oh, actually
I don't see it at all.

1155
00:52:58,530 --> 00:53:01,920
So really, really simulate
that environment accurately.

1156
00:53:01,920 --> 00:53:04,140
If you know your interview
is going to be 45 minutes,

1157
00:53:04,140 --> 00:53:06,150
put a 45-minute timer on your interview.

1158
00:53:06,150 --> 00:53:09,750
If you know the interview is
going to be on a coder pad,

1159
00:53:09,750 --> 00:53:12,990
or it's going to be on a whiteboard,
practice on a whiteboard or pen

1160
00:53:12,990 --> 00:53:14,400
and paper.

1161
00:53:14,400 --> 00:53:17,430
And as you're doing that
practicing, look for patterns.

1162
00:53:17,430 --> 00:53:18,460
Everybody's different.

1163
00:53:18,460 --> 00:53:20,295
Some people are really
great at recursion,

1164
00:53:20,295 --> 00:53:21,670
some people are really bad at it.

1165
00:53:21,670 --> 00:53:23,730
So if you've done 10
practice problems, and you

1166
00:53:23,730 --> 00:53:26,850
found that you got them all right
except for the three recursion ones,

1167
00:53:26,850 --> 00:53:28,800
go practice more recursion.

1168
00:53:28,800 --> 00:53:31,950
Everybody is going to be different in
what they're skilled and not skilled

1169
00:53:31,950 --> 00:53:35,225
at, and so really look for
patterns in your own approach.

1170
00:53:35,225 --> 00:53:37,350
And then lastly, if you
can, practice with friends.

1171
00:53:37,350 --> 00:53:40,200
It can be fun to go
through practice interviews

1172
00:53:40,200 --> 00:53:43,200
with you being the interviewer or
your friend being the interviewer

1173
00:53:43,200 --> 00:53:44,910
and you going through that problem.

1174
00:53:44,910 --> 00:53:47,010
What this does is just
simulates the environment

1175
00:53:47,010 --> 00:53:49,350
and makes it even more
familiar, so that once you're

1176
00:53:49,350 --> 00:53:52,110
in that situation of being
on a real video call,

1177
00:53:52,110 --> 00:53:54,120
writing real code on
a Google doc, it's not

1178
00:53:54,120 --> 00:53:56,578
the first time you've done
that, you're sort of used to it,

1179
00:53:56,578 --> 00:53:59,100
it's really familiar to you.

1180
00:53:59,100 --> 00:54:01,770
So now some tips for
during the interview.

1181
00:54:01,770 --> 00:54:05,280
So all of that was what you should do
in advance, preparing for the interview.

1182
00:54:05,280 --> 00:54:09,110
And so now we're going to jump
into, the Google Meet just started,

1183
00:54:09,110 --> 00:54:12,630
the Zoom call just
started, what do we do?

1184
00:54:12,630 --> 00:54:15,030
So throughout this,
what I really, really

1185
00:54:15,030 --> 00:54:19,380
want to emphasize is that the most
important thing for your interview

1186
00:54:19,380 --> 00:54:22,080
is your process, not your output.

1187
00:54:22,080 --> 00:54:27,000
I've been in many interviews where the
person did not get the correct answer,

1188
00:54:27,000 --> 00:54:31,080
but we hired them, and many
interviews where the person did

1189
00:54:31,080 --> 00:54:33,730
get the correct answer,
and we did not hire them.

1190
00:54:33,730 --> 00:54:37,260
And the reason was, their process
was significantly more important.

1191
00:54:37,260 --> 00:54:40,500
If you can demonstrate that
you're thinking really critically

1192
00:54:40,500 --> 00:54:43,350
and that you're able to
reason through a problem

1193
00:54:43,350 --> 00:54:46,740
and find issues with your
solution and optimize it,

1194
00:54:46,740 --> 00:54:49,440
even if you don't get to a
perfectly correct answer,

1195
00:54:49,440 --> 00:54:53,460
that's so much more important than
somebody who isn't able to communicate,

1196
00:54:53,460 --> 00:54:55,560
who isn't able to think
critically, but is

1197
00:54:55,560 --> 00:54:58,680
able to write correct code just because
they've seen the problem before,

1198
00:54:58,680 --> 00:54:59,430
or they got lucky.

1199
00:54:59,430 --> 00:55:01,138
Or maybe they did
understand the problem,

1200
00:55:01,138 --> 00:55:03,150
but their communication
wasn't there, and so

1201
00:55:03,150 --> 00:55:05,740
it isn't something that you
want to hire onto your team.

1202
00:55:05,740 --> 00:55:07,740
So throughout all this,
really, really emphasize

1203
00:55:07,740 --> 00:55:11,080
that it's not the most important
thing to get the perfect answer,

1204
00:55:11,080 --> 00:55:12,120
but obviously helps.

1205
00:55:12,120 --> 00:55:17,180
But your process is so much more
important than what you produce.

1206
00:55:17,180 --> 00:55:18,920
So the interview will start.

1207
00:55:18,920 --> 00:55:21,620
The interview won't start with
just a problem on the board,

1208
00:55:21,620 --> 00:55:24,590
and someone will say go, but
have the interviewer there.

1209
00:55:24,590 --> 00:55:27,590
They'll probably introduce themselves
and ask you to introduce yourself.

1210
00:55:27,590 --> 00:55:30,465
Something I like to do for the call
starts to just take a deep breath

1211
00:55:30,465 --> 00:55:31,310
and relax.

1212
00:55:31,310 --> 00:55:34,730
Tell yourself that it's going to be OK
and that every interview is stressful

1213
00:55:34,730 --> 00:55:36,500
and everybody hates them.

1214
00:55:36,500 --> 00:55:38,660
So take a deep breath, relax.

1215
00:55:38,660 --> 00:55:41,150
Have an intro prepared,
just like a minute or two,

1216
00:55:41,150 --> 00:55:45,830
like, hi, I'm Tommy, originally
from Boston, worked at Quora for six

1217
00:55:45,830 --> 00:55:48,898
years on back end,
looking for my next role,

1218
00:55:48,898 --> 00:55:50,690
looking to transition
into a frontend role,

1219
00:55:50,690 --> 00:55:52,565
because I want to have
a lot of impact there,

1220
00:55:52,565 --> 00:55:54,890
and I think that space is
really interesting right now.

1221
00:55:54,890 --> 00:55:55,520
Done.

1222
00:55:55,520 --> 00:55:57,562
It doesn't have to be
super fancy, but if someone

1223
00:55:57,562 --> 00:55:59,810
says, hi, introduce yourself,
and you're scrambling,

1224
00:55:59,810 --> 00:56:01,767
and you don't have an
introduction ready,

1225
00:56:01,767 --> 00:56:03,600
it's getting yourself
off on the wrong foot.

1226
00:56:03,600 --> 00:56:05,402
So just have something ready.

1227
00:56:05,402 --> 00:56:08,360
They might also ask you tell me about
the recent project you worked on.

1228
00:56:08,360 --> 00:56:12,590
Have a quick three minutes ready
to go there, but don't ramble.

1229
00:56:12,590 --> 00:56:15,110
This is meant to take
three to five minutes,

1230
00:56:15,110 --> 00:56:16,935
just make sure everybody
feels comfortable.

1231
00:56:16,935 --> 00:56:20,060
So if you're introducing yourself for
30 minutes, which I've had people do,

1232
00:56:20,060 --> 00:56:21,380
you're probably doing it wrong.

1233
00:56:21,380 --> 00:56:23,940


1234
00:56:23,940 --> 00:56:25,800
And then you'll get the problem.

1235
00:56:25,800 --> 00:56:28,790
So after the intros, the
interviewer will give you

1236
00:56:28,790 --> 00:56:31,130
the problem that you're going to solve.

1237
00:56:31,130 --> 00:56:32,430
First thing-- don't panic.

1238
00:56:32,430 --> 00:56:33,680
You're going to get a problem.

1239
00:56:33,680 --> 00:56:35,930
You're probably not going to know the
solution off the top of your head.

1240
00:56:35,930 --> 00:56:36,980
That's expected.

1241
00:56:36,980 --> 00:56:38,600
Don't freak out.

1242
00:56:38,600 --> 00:56:42,980
If you have seen the problem
before, tell your interviewer that.

1243
00:56:42,980 --> 00:56:48,500
So maybe you just practiced some
palindrome problem on LeetCode

1244
00:56:48,500 --> 00:56:51,140
last night, and your interviewer
says, given a string,

1245
00:56:51,140 --> 00:56:52,370
tell me if it's a palindrome.

1246
00:56:52,370 --> 00:56:54,110
Just tell them if you've seen it.

1247
00:56:54,110 --> 00:56:57,590
Or if you've seen something
similar, just say, oh, yeah, this

1248
00:56:57,590 --> 00:57:01,488
was one of the problems I practiced last
week, so I have seen this one before.

1249
00:57:01,488 --> 00:57:03,530
They might say, OK, great,
thanks for telling me.

1250
00:57:03,530 --> 00:57:04,610
Let's solve a different one.

1251
00:57:04,610 --> 00:57:05,810
They might say, OK, that's fine.

1252
00:57:05,810 --> 00:57:06,935
Let's just solve it anyway.

1253
00:57:06,935 --> 00:57:09,660
Even if you've seen it before, I
still want to see your approach.

1254
00:57:09,660 --> 00:57:11,910
It's up to the interviewer
what they want to do there.

1255
00:57:11,910 --> 00:57:15,590
But the thing to not do is be dishonest.

1256
00:57:15,590 --> 00:57:18,320
I catch people all the
time being dishonest.

1257
00:57:18,320 --> 00:57:22,160
One way to get caught is
we sometimes at an on-site,

1258
00:57:22,160 --> 00:57:24,372
two interviewers talking
to the same person

1259
00:57:24,372 --> 00:57:26,330
will ask the same question,
because they forgot

1260
00:57:26,330 --> 00:57:27,650
to share what questions they asked.

1261
00:57:27,650 --> 00:57:29,300
And then the second time,
the person would be like, oh,

1262
00:57:29,300 --> 00:57:31,133
I haven't seen this
before, let me solve it.

1263
00:57:31,133 --> 00:57:33,350
It's just an instant rejection.

1264
00:57:33,350 --> 00:57:35,302
Unless you're an Academy
Award winning actor,

1265
00:57:35,302 --> 00:57:37,010
you'll probably get
caught trying to act.

1266
00:57:37,010 --> 00:57:38,270
So just be honest.

1267
00:57:38,270 --> 00:57:41,150
Don't try to do anything sneaky here.

1268
00:57:41,150 --> 00:57:44,210
Just let your interviewer know
if you've seen the problem.

1269
00:57:44,210 --> 00:57:46,280
After you get the
problem-- let's say you've

1270
00:57:46,280 --> 00:57:47,867
disclosed if you've seen it before.

1271
00:57:47,867 --> 00:57:51,200
After that, the first thing you probably
want to do is ask clarifying questions.

1272
00:57:51,200 --> 00:57:56,010
Really good clarifying questions are of
the form, OK, so let me just be clear.

1273
00:57:56,010 --> 00:57:58,490
So if this was the input,
then this would be my output?

1274
00:57:58,490 --> 00:58:00,070
And they'll say, yes, it is.

1275
00:58:00,070 --> 00:58:00,920
You say, OK, great.

1276
00:58:00,920 --> 00:58:05,100
You're on the same page,
you've understood the problem.

1277
00:58:05,100 --> 00:58:08,223
Then as you're solving
it, always think out loud.

1278
00:58:08,223 --> 00:58:10,140
You should basically,
as an interviewer, never

1279
00:58:10,140 --> 00:58:15,270
be silent for more than like 30
seconds or 60 seconds at a time.

1280
00:58:15,270 --> 00:58:18,480
It's going to feel really weird to
constantly just be thinking out loud,

1281
00:58:18,480 --> 00:58:22,170
because that's not how most developers
or day scientists or designers think.

1282
00:58:22,170 --> 00:58:25,230
They just think in their
head and not speaking.

1283
00:58:25,230 --> 00:58:27,300
Your interviewer will
expect you to do this.

1284
00:58:27,300 --> 00:58:29,590
It will feel weird if
you've never done it before,

1285
00:58:29,590 --> 00:58:32,418
but it is what you're supposed to do.

1286
00:58:32,418 --> 00:58:35,460
Sometimes you might say, OK, do you
mind if I just take a minute to think

1287
00:58:35,460 --> 00:58:37,100
or draw something out
on a piece of paper?

1288
00:58:37,100 --> 00:58:38,392
The interviewer will say, sure.

1289
00:58:38,392 --> 00:58:41,208
But if you're sitting there
in silence for five minutes

1290
00:58:41,208 --> 00:58:43,500
typing away at your keyboard
and not making much sense,

1291
00:58:43,500 --> 00:58:46,110
your interview is not going well.

1292
00:58:46,110 --> 00:58:49,620
If you can, try talking through
multiple possible approaches.

1293
00:58:49,620 --> 00:58:51,960
Usually when you're given
a problem, you might have

1294
00:58:51,960 --> 00:58:53,962
a couple ideas for how to solve it.

1295
00:58:53,962 --> 00:58:55,170
And talk through all of them.

1296
00:58:55,170 --> 00:58:57,003
You can say, OK, I have
a couple ideas here.

1297
00:58:57,003 --> 00:59:00,780
One idea is to build a graph
that has these as nodes.

1298
00:59:00,780 --> 00:59:03,720
Another idea would be to create a
hash table, where this is the key,

1299
00:59:03,720 --> 00:59:04,800
and this is the value.

1300
00:59:04,800 --> 00:59:06,300
Sort of explain it out loud.

1301
00:59:06,300 --> 00:59:09,660
Your interviewer is probably going
to help you here and say, OK, yeah,

1302
00:59:09,660 --> 00:59:10,753
those both might work.

1303
00:59:10,753 --> 00:59:12,420
If I were you, I'd go at the second one.

1304
00:59:12,420 --> 00:59:14,820
You can say, OK, great, now I'm
going to go with the second one.

1305
00:59:14,820 --> 00:59:16,800
You're sort of making it
more of a conversation

1306
00:59:16,800 --> 00:59:20,880
and giving your interviewer
an opportunity to help you.

1307
00:59:20,880 --> 00:59:25,170
Many folks are sort of visual folks,
so feel free to draw a diagram.

1308
00:59:25,170 --> 00:59:27,120
Even if you're doing
this over video, having

1309
00:59:27,120 --> 00:59:29,250
a little whiteboard or
a piece of paper handy,

1310
00:59:29,250 --> 00:59:31,300
that can be helpful if
you're a visual person.

1311
00:59:31,300 --> 00:59:35,020
So if you want to draw the diagram,
hold it up to the webcam if you want,

1312
00:59:35,020 --> 00:59:37,710
or not, depending on
what the situation is.

1313
00:59:37,710 --> 00:59:42,030
But have the supplies ready
if you think you'll need them.

1314
00:59:42,030 --> 00:59:43,860
Then after you've got
the problem, you've

1315
00:59:43,860 --> 00:59:46,890
articulated some solutions,
what you really want to do

1316
00:59:46,890 --> 00:59:49,410
is try to pattern match.

1317
00:59:49,410 --> 00:59:53,160
We mentioned that most
interview problems reduce down

1318
00:59:53,160 --> 00:59:55,740
into a really small
number of core concepts

1319
00:59:55,740 --> 00:59:58,950
like search problems,
strings, graph problems.

1320
00:59:58,950 --> 01:00:03,238
So try to pattern match
against what you know to what

1321
01:00:03,238 --> 01:00:04,780
you think this problem is getting at.

1322
01:00:04,780 --> 01:00:07,320
So for instance, maybe someone's
described something where

1323
01:00:07,320 --> 01:00:09,487
you're searching for
something that could

1324
01:00:09,487 --> 01:00:11,820
be a bunch of different
possibilities, ask yourself, OK,

1325
01:00:11,820 --> 01:00:15,060
could we formulate that as a search?

1326
01:00:15,060 --> 01:00:18,060
Or maybe somebody has described some
problem, and you're like, OK, well,

1327
01:00:18,060 --> 01:00:21,150
maybe there's some recursion
here that I can formulate,

1328
01:00:21,150 --> 01:00:23,010
or maybe binary search is applicable.

1329
01:00:23,010 --> 01:00:26,400
But take all those things we talked
about earlier in your toolbox

1330
01:00:26,400 --> 01:00:31,320
like lists, strings, trees, hash tables,
and just go through them in your head.

1331
01:00:31,320 --> 01:00:32,580
Does a hash table apply here?

1332
01:00:32,580 --> 01:00:34,050
Does a list apply here?

1333
01:00:34,050 --> 01:00:35,745
Does a tree apply here?

1334
01:00:35,745 --> 01:00:37,620
Hopefully, the answer
to one of those is yes,

1335
01:00:37,620 --> 01:00:42,000
and that can really help kick start your
solution and get you off the ground.

1336
01:00:42,000 --> 01:00:45,420
And then you sort of figure
out, why is it a graph problem,

1337
01:00:45,420 --> 01:00:47,670
or how is this different
than a standard graph search?

1338
01:00:47,670 --> 01:00:51,390
But it can help get you to a
starting point for your solution.

1339
01:00:51,390 --> 01:00:55,302
Rather than trying to invent some
new algorithm on the spot or derive

1340
01:00:55,302 --> 01:00:58,260
something totally new in the spot,
pattern match against what you know.

1341
01:00:58,260 --> 01:01:01,427
And you know a lot, because you went
through all of the toolbox and prep you

1342
01:01:01,427 --> 01:01:04,170
should be doing in advance.

1343
01:01:04,170 --> 01:01:09,060
As you're writing up your
solution, try to write general code

1344
01:01:09,060 --> 01:01:10,470
before really specific code.

1345
01:01:10,470 --> 01:01:15,630
And what I mean by that is, let's say
that your interview problem is, write

1346
01:01:15,630 --> 01:01:17,340
a function to compute a factorial.

1347
01:01:17,340 --> 01:01:19,780
What some people do as their
first instinct is they say,

1348
01:01:19,780 --> 01:01:23,280
well, I know the factorial of 3 is 6.

1349
01:01:23,280 --> 01:01:25,980
So I'm going to say if
input equals 3, return 6.

1350
01:01:25,980 --> 01:01:28,770
And then they come up with like
three or four corner cases first.

1351
01:01:28,770 --> 01:01:31,080
I would say, don't do that.

1352
01:01:31,080 --> 01:01:33,060
That code is probably
not going to last--

1353
01:01:33,060 --> 01:01:34,680
you're going to end up deleting it.

1354
01:01:34,680 --> 01:01:37,620
But it can also sometimes
create this mental block

1355
01:01:37,620 --> 01:01:40,260
where you fail to see
this general solution,

1356
01:01:40,260 --> 01:01:43,930
because you're zoomed in on a small
number of cases you know work.

1357
01:01:43,930 --> 01:01:48,030
So try to think about this generalized
solution, and then find corner cases,

1358
01:01:48,030 --> 01:01:50,370
rather than enumerate all
the possible corner cases

1359
01:01:50,370 --> 01:01:53,890
and then try to find
some general solution.

1360
01:01:53,890 --> 01:01:58,210
For the most part, too, if it feels like
your solution is really complicated,

1361
01:01:58,210 --> 01:02:01,900
like it's got a bunch of corner cases
and a bunch of branching and four

1362
01:02:01,900 --> 01:02:05,260
different helper functions, if
it just feels pretty complicated,

1363
01:02:05,260 --> 01:02:09,502
it probably is, and there
probably is a simpler solution.

1364
01:02:09,502 --> 01:02:11,710
Sometimes there's not, and
it's kind of hard to tell,

1365
01:02:11,710 --> 01:02:15,010
but your interviewer is probably going
to be there asking you to simplify it,

1366
01:02:15,010 --> 01:02:16,360
or maybe there's a better way.

1367
01:02:16,360 --> 01:02:19,552
But just as a heuristic for yourself,
if you're looking at your code

1368
01:02:19,552 --> 01:02:22,010
and you're like, wow, that's
a lot, it feels really messy--

1369
01:02:22,010 --> 01:02:24,135
it's not because like, oh,
the problem is so hard--

1370
01:02:24,135 --> 01:02:27,860
it's probably because your
solution is overly complicated.

1371
01:02:27,860 --> 01:02:31,440
To get out a question that we
asked earlier, if you get stuck,

1372
01:02:31,440 --> 01:02:32,690
ask for help.

1373
01:02:32,690 --> 01:02:35,992
It's much better to say, honestly,
I'm a little bit stuck here.

1374
01:02:35,992 --> 01:02:36,950
Can you give me a hint?

1375
01:02:36,950 --> 01:02:38,742
Or can you point me in
the right direction?

1376
01:02:38,742 --> 01:02:42,590
Saying that is so much better than
just staring at a wall for 30 minutes

1377
01:02:42,590 --> 01:02:44,330
and getting nowhere on the problem.

1378
01:02:44,330 --> 01:02:47,390
It's obviously better if you solve
the problem without asking for hints.

1379
01:02:47,390 --> 01:02:50,450
Someone who's constantly asking for
hints and needs a lot of hand-holding

1380
01:02:50,450 --> 01:02:51,380
to get to the answer--

1381
01:02:51,380 --> 01:02:53,120
that's not a good interview either.

1382
01:02:53,120 --> 01:02:55,160
But sometimes you do hit a wall.

1383
01:02:55,160 --> 01:02:58,490
You can ask a question, and with a
small little hint, you can get unstuck,

1384
01:02:58,490 --> 01:03:01,860
and then you're back off to the races.

1385
01:03:01,860 --> 01:03:04,710
And also make sure before you
type a single line of code,

1386
01:03:04,710 --> 01:03:05,850
explain your solution.

1387
01:03:05,850 --> 01:03:08,490
Be aligned on what you're about
to write before you write it.

1388
01:03:08,490 --> 01:03:11,850
You don't want to create some situation
where you've written a bunch of code

1389
01:03:11,850 --> 01:03:12,900
in the back of your
head, your interviewer

1390
01:03:12,900 --> 01:03:15,330
is like, they didn't even tell
me what they were trying to do.

1391
01:03:15,330 --> 01:03:17,730
I don't even know what they're trying
to do, and this is super wrong,

1392
01:03:17,730 --> 01:03:18,930
and you don't have
enough time to fix it.

1393
01:03:18,930 --> 01:03:20,640
And then you interview's over.

1394
01:03:20,640 --> 01:03:21,450
You lose.

1395
01:03:21,450 --> 01:03:25,020
So make sure you're aligned on the same
page of what code you're about to write

1396
01:03:25,020 --> 01:03:29,460
and what your general approach
is before you go and write it.

1397
01:03:29,460 --> 01:03:33,160
And again, this is what I mean
when I say process, not output.

1398
01:03:33,160 --> 01:03:36,010
If you're aligned in an approach,
and you're thinking critically,

1399
01:03:36,010 --> 01:03:38,860
and you're communicating well,
maybe your output isn't perfect,

1400
01:03:38,860 --> 01:03:43,170
but that's what we're really
looking for for a good interview.

1401
01:03:43,170 --> 01:03:47,260
OK, some few last tips before we
get into the mock interviews here.

1402
01:03:47,260 --> 01:03:50,700
So as you're writing code,
constantly communicate.

1403
01:03:50,700 --> 01:03:53,940
Sometimes I'll even kind of
narrate the code I'm writing.

1404
01:03:53,940 --> 01:03:57,525
I'll say, well, first, I'm going
to create a function called Cache.

1405
01:03:57,525 --> 01:03:59,400
And then that's going
to take two parameters.

1406
01:03:59,400 --> 01:04:02,358
It's going to take the key, and then
it can take the size of the cache.

1407
01:04:02,358 --> 01:04:05,980
And the first thing I want to do is
I'm going to create a base case here.

1408
01:04:05,980 --> 01:04:09,000
That kind of communication is
really helpful as you're typing.

1409
01:04:09,000 --> 01:04:10,800
Again, it feels weird, it's unnatural.

1410
01:04:10,800 --> 01:04:11,520
Practice it.

1411
01:04:11,520 --> 01:04:14,820
I promise that's what your
interviewer wants you to be doing.

1412
01:04:14,820 --> 01:04:18,420
Next thing is always
get something working.

1413
01:04:18,420 --> 01:04:22,860
Sometimes people will see
some simple or naive solution,

1414
01:04:22,860 --> 01:04:26,007
they know it's not optimal, and rather
than writing up the naive solution,

1415
01:04:26,007 --> 01:04:27,840
they sit there and
they're like, no, I don't

1416
01:04:27,840 --> 01:04:30,548
want to write anything until I
get the perfect, optimal solution.

1417
01:04:30,548 --> 01:04:34,242
Then they end the interview having
written zero lines of code--

1418
01:04:34,242 --> 01:04:36,450
really good chance that the
interview didn't go well.

1419
01:04:36,450 --> 01:04:37,825
You're not going to move forward.

1420
01:04:37,825 --> 01:04:41,250
It's definitely better to have
something than to have nothing.

1421
01:04:41,250 --> 01:04:45,600
Once you have something, then
go ahead and try to simplify,

1422
01:04:45,600 --> 01:04:48,540
make optimizations, and get
to that optimal solution.

1423
01:04:48,540 --> 01:04:51,150
But even if you can't ultimately
get to the optimal solution,

1424
01:04:51,150 --> 01:04:52,530
but you're thinking
really quickly, you're

1425
01:04:52,530 --> 01:04:55,740
communicating well, maybe with five more
minutes you would have gotten there,

1426
01:04:55,740 --> 01:04:58,320
that's a really good interview outcome.

1427
01:04:58,320 --> 01:05:01,410
Last, listen to your interviewer.

1428
01:05:01,410 --> 01:05:04,680
I've had many candidates
who do not listen to me

1429
01:05:04,680 --> 01:05:06,750
as I'm trying to help and give hints.

1430
01:05:06,750 --> 01:05:08,730
Your interviewer wants you to succeed.

1431
01:05:08,730 --> 01:05:12,763
As an interviewer, it's so much
less fun watching somebody struggle,

1432
01:05:12,763 --> 01:05:15,930
and you're trying to help, and they're
telling you, no, you don't understand

1433
01:05:15,930 --> 01:05:17,138
the problem you just gave me.

1434
01:05:17,138 --> 01:05:19,870
It's like, I probably do,
and I want you to succeed.

1435
01:05:19,870 --> 01:05:22,020
So if your interviewer is
giving you little hints

1436
01:05:22,020 --> 01:05:26,160
and guiding you in the right direction,
really listen to what they're saying

1437
01:05:26,160 --> 01:05:29,310
and take it seriously.

1438
01:05:29,310 --> 01:05:31,320
It's really, really rare
that some interviewer

1439
01:05:31,320 --> 01:05:34,080
is going to intentionally try to
fool you or give you some hint

1440
01:05:34,080 --> 01:05:36,750
to lead you down the wrong
direction and be like, oh,

1441
01:05:36,750 --> 01:05:39,702
is that your final answer,
when it's actually correct.

1442
01:05:39,702 --> 01:05:40,410
It's pretty rare.

1443
01:05:40,410 --> 01:05:43,617
I assume the interviewer is not doing
that, unless you know they've already

1444
01:05:43,617 --> 01:05:44,700
done it once or something.

1445
01:05:44,700 --> 01:05:46,110
So really listen to your interviewer.

1446
01:05:46,110 --> 01:05:47,130
They want to help you.

1447
01:05:47,130 --> 01:05:51,880
Take advantage of that and
listen to their feedback.

1448
01:05:51,880 --> 01:05:54,670
Oftentimes people ask,
what code do I write?

1449
01:05:54,670 --> 01:05:59,390
What we're looking for specifically
is a function that returns the answer.

1450
01:05:59,390 --> 01:06:01,600
So don't just write a linear--

1451
01:06:01,600 --> 01:06:04,180
the first line of code
shouldn't be like, if this,

1452
01:06:04,180 --> 01:06:05,740
and assume some variable exists.

1453
01:06:05,740 --> 01:06:07,990
The first line you write
should be creating a function

1454
01:06:07,990 --> 01:06:09,850
that will ultimately return the answer.

1455
01:06:09,850 --> 01:06:12,940
If that function needs to be
more than one function, great.

1456
01:06:12,940 --> 01:06:15,680
Decompose that into separate
functions as needed.

1457
01:06:15,680 --> 01:06:16,648
Don't go crazy here.

1458
01:06:16,648 --> 01:06:19,690
And make sure you have something
working before you spend a bunch of time

1459
01:06:19,690 --> 01:06:22,840
refactoring code and cleaning it up.

1460
01:06:22,840 --> 01:06:27,340
A lot of good general engineering and
programming principles apply here too.

1461
01:06:27,340 --> 01:06:30,610
Don't create variable names with
crazy acronyms or abbreviations.

1462
01:06:30,610 --> 01:06:32,170
Make sure everything is readable.

1463
01:06:32,170 --> 01:06:37,090
If you have common logic, factor
things out into helper functions.

1464
01:06:37,090 --> 01:06:39,430
Again, this is all about the process.

1465
01:06:39,430 --> 01:06:41,830
If you're doing this and
demonstrating these things,

1466
01:06:41,830 --> 01:06:46,860
that's way more important than your
function being perfect at the end.

1467
01:06:46,860 --> 01:06:48,600
Last part is testing it.

1468
01:06:48,600 --> 01:06:53,100
Once you have written code, your
job now is to be a computer.

1469
01:06:53,100 --> 01:06:56,280
So take your code and run
through it line-by-line.

1470
01:06:56,280 --> 01:06:59,010
Execute the code as if you were a CPU.

1471
01:06:59,010 --> 01:07:01,170
Make sure that if you
run the code, it's going

1472
01:07:01,170 --> 01:07:03,100
to end up returning the correct answer.

1473
01:07:03,100 --> 01:07:04,870
And again, do that out loud.

1474
01:07:04,870 --> 01:07:07,953
So what I would say is, let's say you
have that factorial function saying,

1475
01:07:07,953 --> 01:07:09,360
OK, well, let's try passing in 3.

1476
01:07:09,360 --> 01:07:10,170
Well, it's 3.

1477
01:07:10,170 --> 01:07:12,150
My base case is checking
if it's equal to 1.

1478
01:07:12,150 --> 01:07:13,233
3 is not equal to 1.

1479
01:07:13,233 --> 01:07:14,650
So let's move on to the next line.

1480
01:07:14,650 --> 01:07:16,358
We're going to enter
this recursive case.

1481
01:07:16,358 --> 01:07:19,890
And just walk it through and make
sure that you get the correct answer

1482
01:07:19,890 --> 01:07:21,150
at the end of it.

1483
01:07:21,150 --> 01:07:23,550
As you're doing this, be
on the lookout for bugs.

1484
01:07:23,550 --> 01:07:27,060
Sometimes in your head,
what the code is doing

1485
01:07:27,060 --> 01:07:28,810
is different than
what's actually written.

1486
01:07:28,810 --> 01:07:32,040
So as best you can, do exactly
what the code is actually doing,

1487
01:07:32,040 --> 01:07:33,690
and that's going to help you find bugs.

1488
01:07:33,690 --> 01:07:37,080
And just verbally reason about
that out loud and walk through,

1489
01:07:37,080 --> 01:07:42,370
sort of narrate the process of this
function executing to your interviewer.

1490
01:07:42,370 --> 01:07:45,970
Sometimes in an interview, you'll
be given a computer and the ability

1491
01:07:45,970 --> 01:07:46,750
to run code.

1492
01:07:46,750 --> 01:07:49,270
Sometimes you won't, either
because it's on a whiteboard,

1493
01:07:49,270 --> 01:07:51,610
or because the interviewer says
you're not allowed to run code.

1494
01:07:51,610 --> 01:07:53,140
I want to see your reason about it.

1495
01:07:53,140 --> 01:07:54,848
In the mock interviews
we're about to do,

1496
01:07:54,848 --> 01:07:57,350
we will be allowed to
run code on the computer.

1497
01:07:57,350 --> 01:08:01,252
So if you can do that, I advise to
write a few test cases, if you can.

1498
01:08:01,252 --> 01:08:04,210
Maybe they're just like print statements
at the end or something that's

1499
01:08:04,210 --> 01:08:07,510
going to run the code and
verify the output is correct.

1500
01:08:07,510 --> 01:08:11,260
But what I wouldn't do when you're
given the ability to run the code

1501
01:08:11,260 --> 01:08:14,740
is take advantage of that just
by running it a million times.

1502
01:08:14,740 --> 01:08:17,380
And let's say you have an
off by 1 error somewhere,

1503
01:08:17,380 --> 01:08:20,067
just randomly adding plus
1's and minus 1's and then

1504
01:08:20,067 --> 01:08:22,359
running it until eventually
you get the correct answer.

1505
01:08:22,359 --> 01:08:24,567
Even if the answer is correct,
that's a great example

1506
01:08:24,567 --> 01:08:27,130
of a process that is not good.

1507
01:08:27,130 --> 01:08:30,220
We want to see reasoning about
things and arriving at a solution

1508
01:08:30,220 --> 01:08:34,060
confidently, rather than just randomly
locking your way into a solution.

1509
01:08:34,060 --> 01:08:37,300
And the ability to run code a
bunch can be this temptation that

1510
01:08:37,300 --> 01:08:39,970
can lead to that kind of outcome.

1511
01:08:39,970 --> 01:08:43,000
And last, if you can,
add print statements

1512
01:08:43,000 --> 01:08:45,520
or debug the state as
your code is executing.

1513
01:08:45,520 --> 01:08:46,550
Definitely do that.

1514
01:08:46,550 --> 01:08:48,850
I think sometimes people will
run the code, look at the output,

1515
01:08:48,850 --> 01:08:51,425
say that's wrong, change
something, look at just the output,

1516
01:08:51,425 --> 01:08:52,300
and say that's wrong.

1517
01:08:52,300 --> 01:08:54,819
We're printing out intermediate
state, and what your code

1518
01:08:54,819 --> 01:08:56,229
is doing along the way.

1519
01:08:56,229 --> 01:08:57,819
Maybe you have a recursive function.

1520
01:08:57,819 --> 01:09:00,950
You're going to print every time the
recursion is called, print the inputs.

1521
01:09:00,950 --> 01:09:04,130
You can make sure the stack is what
you think it's going to look like.

1522
01:09:04,130 --> 01:09:07,580
And that's usually the most
effective way to fix issues.

1523
01:09:07,580 --> 01:09:09,410
I think I'll say it one last time--

1524
01:09:09,410 --> 01:09:13,460
the process matters way
more than the output here.

1525
01:09:13,460 --> 01:09:16,760
OK, so let's jump into
some mock interviews.

1526
01:09:16,760 --> 01:09:21,740
So what we'll do is we'll have Connor,
and Maria will be the interviewee.

1527
01:09:21,740 --> 01:09:23,840
I will be the interviewer.

1528
01:09:23,840 --> 01:09:28,694
And we'll try to simulate this
really as realistically as we can.

1529
01:09:28,694 --> 01:09:31,069
So like I said, these are real
problems I've really given

1530
01:09:31,069 --> 01:09:33,082
in the past at real tech companies.

1531
01:09:33,082 --> 01:09:35,540
And then Connor or Maria have
done lots of real interviews,

1532
01:09:35,540 --> 01:09:38,498
and so they'll give you what a real
performance looks like, just so you

1533
01:09:38,498 --> 01:09:40,740
can take a look at how things go.

1534
01:09:40,740 --> 01:09:44,120
So I'll read the problem first, and
then we'll flip into screen sharing

1535
01:09:44,120 --> 01:09:45,930
and start the interview.

1536
01:09:45,930 --> 01:09:49,017
So this is the first problem that
we're going to be working on.

1537
01:09:49,017 --> 01:09:49,850
So I'll read it out.

1538
01:09:49,850 --> 01:09:52,670
So let's define a rotated
array as a sorted array

1539
01:09:52,670 --> 01:09:56,810
where the numbers have all been rotated
to the right some number of places,

1540
01:09:56,810 --> 01:09:59,220
with numbers wrapping around
when they reach the end.

1541
01:09:59,220 --> 01:10:02,030
So for instance, if you've
got the array, 1, 2, 3, 4, 5,

1542
01:10:02,030 --> 01:10:06,740
and you rotate that three times,
you're going to get 3, 4, 5, 1, 2.

1543
01:10:06,740 --> 01:10:10,010
So the problem here is
given, a rotated array,

1544
01:10:10,010 --> 01:10:12,877
find the number of times it was rotated.

1545
01:10:12,877 --> 01:10:13,960
CONNOR LEGGETT: All right.

1546
01:10:13,960 --> 01:10:14,710
Well, hi, Tommy.

1547
01:10:14,710 --> 01:10:15,460
How are you doing?

1548
01:10:15,460 --> 01:10:16,060
I'm Connor.

1549
01:10:16,060 --> 01:10:17,110
TOMMY MACWILLIAMS: Good,
nice to meet you Connor.

1550
01:10:17,110 --> 01:10:18,247
How's your day going?

1551
01:10:18,247 --> 01:10:20,330
CONNOR LEGGETT: It's going
pretty well, and yours?

1552
01:10:20,330 --> 01:10:21,310
TOMMY MACWILLIAMS: Awesome, not too bad.

1553
01:10:21,310 --> 01:10:23,630
It's pretty sunny out
here in San Francisco,

1554
01:10:23,630 --> 01:10:25,732
which doesn't happen too often.

1555
01:10:25,732 --> 01:10:26,940
CONNOR LEGGETT: That's great.

1556
01:10:26,940 --> 01:10:30,500
Yeah, it's pretty warm out
here for October in Cambridge.

1557
01:10:30,500 --> 01:10:31,910
DAVID MALAN: Nice.

1558
01:10:31,910 --> 01:10:34,140
All right, well, thanks
for taking the time.

1559
01:10:34,140 --> 01:10:35,760
Let's jump right into the question.

1560
01:10:35,760 --> 01:10:38,360
So you've got the Replit
open, that's great.

1561
01:10:38,360 --> 01:10:40,640
I notice that you've
selected Python there.

1562
01:10:40,640 --> 01:10:43,070
Is that the language that
you want to use for this?

1563
01:10:43,070 --> 01:10:44,195
CONNOR LEGGETT: Yes, it is.

1564
01:10:44,195 --> 01:10:46,800
TOMMY MACWILLIAMS:
Perfect, that works for me.

1565
01:10:46,800 --> 01:10:49,770
So here's the question
that I'd like you to solve.

1566
01:10:49,770 --> 01:10:54,590
So let's define a rotated array as a
sorted array, where the numbers have

1567
01:10:54,590 --> 01:10:57,710
all been rotated to the
right some number of places,

1568
01:10:57,710 --> 01:11:00,270
with numbers wrapping around
when they reach the end.

1569
01:11:00,270 --> 01:11:03,080
So for instance, if I give
you the input 1, 2, 3, 4,

1570
01:11:03,080 --> 01:11:06,340
5, you should return back 3, since--

1571
01:11:06,340 --> 01:11:08,090
oh, sorry, when I
rotate that three times,

1572
01:11:08,090 --> 01:11:15,780
you should return back 3, basically,
as you've written in the comment there.

1573
01:11:15,780 --> 01:11:19,050
So that's the question
that I'd like you to solve.

1574
01:11:19,050 --> 01:11:21,360
CONNOR LEGGETT: OK,
yeah, that sounds good.

1575
01:11:21,360 --> 01:11:24,285
So I'll start out by asking
some clarifying questions,

1576
01:11:24,285 --> 01:11:25,410
if that sounds good to you.

1577
01:11:25,410 --> 01:11:26,730
TOMMY MACWILLIAMS: Yeah.

1578
01:11:26,730 --> 01:11:29,980
CONNOR LEGGETT: So just to make sure,
it does say, given a rotated array here.

1579
01:11:29,980 --> 01:11:32,550
I just want to clarify that
something like this right

1580
01:11:32,550 --> 01:11:35,140
here would be our input.

1581
01:11:35,140 --> 01:11:38,010
So I don't get this
original array at all,

1582
01:11:38,010 --> 01:11:39,870
I just get the rotated array as input?

1583
01:11:39,870 --> 01:11:41,900
TOMMY MACWILLIAMS: Yep, that's right.

1584
01:11:41,900 --> 01:11:44,870
CONNOR LEGGETT: All right, and
then my other clarifying question

1585
01:11:44,870 --> 01:11:52,992
is, in this one, all of these are
consecutive integers, so 1, 2, 3, 4, 5,

1586
01:11:52,992 --> 01:11:54,200
all right next to each other.

1587
01:11:54,200 --> 01:11:57,080
Can I assume that that will
be the case, or might there

1588
01:11:57,080 --> 01:12:04,635
be a case where my original
array is like 1, 3, 4, 7, 10?

1589
01:12:04,635 --> 01:12:06,510
TOMMY MACWILLIAMS: Yeah,
so that could exist.

1590
01:12:06,510 --> 01:12:09,918
So they're not necessarily
consecutive, but they are sorted.

1591
01:12:09,918 --> 01:12:10,710
CONNOR LEGGETT: OK.

1592
01:12:10,710 --> 01:12:12,480
Yeah, that's good to know.

1593
01:12:12,480 --> 01:12:14,220
All right.

1594
01:12:14,220 --> 01:12:19,350
In that case, let me just
look at a couple of examples

1595
01:12:19,350 --> 01:12:22,110
just so I can see if we can
start to see a pattern here.

1596
01:12:22,110 --> 01:12:23,110
TOMMY MACWILLIAMS: Sure.

1597
01:12:23,110 --> 01:12:26,080
CONNOR LEGGETT: So we see when
we get this first example,

1598
01:12:26,080 --> 01:12:32,850
we want that to map to 3, because
we're rotating that one 3 times.

1599
01:12:32,850 --> 01:12:37,000
Let me see if I were to
rotate that one more time,

1600
01:12:37,000 --> 01:12:42,030
then it would look like 2, 3, 4, 5, 1.

1601
01:12:42,030 --> 01:12:47,520
And so that would be as if
I had rotated it 4 times.

1602
01:12:47,520 --> 01:12:50,110
And then I'll try some ones
with different numbers.

1603
01:12:50,110 --> 01:12:57,990
So if my original array has
some different numbers in it,

1604
01:12:57,990 --> 01:13:02,490
a rotated one might look
like 6, 8, 12, 1, 3.

1605
01:13:02,490 --> 01:13:08,120


1606
01:13:08,120 --> 01:13:11,090
In this case, we would
have rotated it so one

1607
01:13:11,090 --> 01:13:13,340
would have been the original one.

1608
01:13:13,340 --> 01:13:15,900
And so we've rotated it 1, 2, 3 times.

1609
01:13:15,900 --> 01:13:20,300
So we would also want
this one to return 3.

1610
01:13:20,300 --> 01:13:27,720
And then just one more kind of edge-ish
case is if it's not sorted at all.

1611
01:13:27,720 --> 01:13:31,640
I'm assuming that we would want to
say that that was rotated 0 times.

1612
01:13:31,640 --> 01:13:33,470
TOMMY MACWILLIAMS: Yes, that's right.

1613
01:13:33,470 --> 01:13:34,553
CONNOR LEGGETT: All right.

1614
01:13:34,553 --> 01:13:37,400
So just to check in, do all of
those cases seem good to you?

1615
01:13:37,400 --> 01:13:40,830
TOMMY MACWILLIAMS: Yeah, nothing
else you'll probably run into here.

1616
01:13:40,830 --> 01:13:42,380
CONNOR LEGGETT: All right, great.

1617
01:13:42,380 --> 01:13:45,390
So in that case, let me go ahead and
run through, looking for patterns.

1618
01:13:45,390 --> 01:13:48,620
So I can see just how
I would figure this out

1619
01:13:48,620 --> 01:13:52,880
if I was looking through this as a
human, is I would go to the first one,

1620
01:13:52,880 --> 01:13:56,960
and I guess the important thing I'm
looking for is all of these numbers

1621
01:13:56,960 --> 01:13:58,160
are increasing.

1622
01:13:58,160 --> 01:14:02,190
And I want to look for the
first time that they decrease,

1623
01:14:02,190 --> 01:14:03,650
so they start going down.

1624
01:14:03,650 --> 01:14:09,110
So in this case, I go to 0,
I go to 1, it's increasing.

1625
01:14:09,110 --> 01:14:10,820
I go to 2, it's increasing.

1626
01:14:10,820 --> 01:14:16,050
And then when I get to
element 3, it's decreasing,

1627
01:14:16,050 --> 01:14:19,220
which means it's rotated 3 times.

1628
01:14:19,220 --> 01:14:22,590
OK, and then just to verify
that, I'll go through this one.

1629
01:14:22,590 --> 01:14:27,260
So in this one, I started element
0, then 1, it's still increasing,

1630
01:14:27,260 --> 01:14:29,400
2, still increasing,
3, still increasing.

1631
01:14:29,400 --> 01:14:32,060
But at 4 it decreases again.

1632
01:14:32,060 --> 01:14:35,960
So I think that is a pretty
good pattern to use there,

1633
01:14:35,960 --> 01:14:40,770
is that when we get to an index
when it starts decreasing,

1634
01:14:40,770 --> 01:14:42,807
that's the index that we want to return.

1635
01:14:42,807 --> 01:14:45,140
TOMMY MACWILLIAMS: Yeah, that
looks exactly right to me.

1636
01:14:45,140 --> 01:14:48,223
Yeah, that point where it goes from
increasing to decreasing is the index.

1637
01:14:48,223 --> 01:14:50,672
That tells you the number of
times it had been rotated.

1638
01:14:50,672 --> 01:14:52,130
CONNOR LEGGETT: All right, perfect.

1639
01:14:52,130 --> 01:14:54,500
In that case, I think I'll
go into writing some code,

1640
01:14:54,500 --> 01:14:55,400
if that sounds good to you.

1641
01:14:55,400 --> 01:14:57,020
TOMMY MACWILLIAMS: Yeah, let's see that.

1642
01:14:57,020 --> 01:14:58,103
CONNOR LEGGETT: All right.

1643
01:14:58,103 --> 01:15:01,350
So I'll go ahead and
define a function here.

1644
01:15:01,350 --> 01:15:07,070
And so I guess I'll just
say, get rotate number.

1645
01:15:07,070 --> 01:15:09,280
TOMMY MACWILLIAMS: Sure, sounds good.

1646
01:15:09,280 --> 01:15:14,230
CONNOR LEGGETT: And it's
going to take in this array.

1647
01:15:14,230 --> 01:15:20,080
Ah, that looks like it might be
reserved, so I'll take in a list, ls.

1648
01:15:20,080 --> 01:15:25,270
Now I want to go through
all of the elements in ls.

1649
01:15:25,270 --> 01:15:27,520
And the index is important
here, so I guess I'll

1650
01:15:27,520 --> 01:15:32,590
go ahead and say for i in range ls.

1651
01:15:32,590 --> 01:15:37,120
And now what I want to do is for
each i-- like when I'm here, i is 2--

1652
01:15:37,120 --> 01:15:42,190
I want to compare it to the one before
that and see if it's less than that.

1653
01:15:42,190 --> 01:15:43,320
So let me--

1654
01:15:43,320 --> 01:15:47,140
I guess I would say
something like, if ls bracket

1655
01:15:47,140 --> 01:15:52,720
i is less than ls bracket i minus 1--

1656
01:15:52,720 --> 01:15:59,170
so we're saying, if my current element
is less than the element before that--

1657
01:15:59,170 --> 01:16:04,510
then I would go ahead, and I know that
that's the place that I want to stop,

1658
01:16:04,510 --> 01:16:08,740
so I can go ahead and say return i.

1659
01:16:08,740 --> 01:16:13,120
And now, looking at this special
case here where nothing happens,

1660
01:16:13,120 --> 01:16:16,970
I guess I would go through this entire
list and never get to that point.

1661
01:16:16,970 --> 01:16:22,480
So I guess if I finish this whole
thing, so I don't find any point where

1662
01:16:22,480 --> 01:16:27,100
an element is less than its predecessor,
I'm going to go ahead and return 0,

1663
01:16:27,100 --> 01:16:32,352
since I know that the list
is not rotated at all.

1664
01:16:32,352 --> 01:16:34,060
So now, just as a
sanity check, I'm going

1665
01:16:34,060 --> 01:16:38,540
to go through this with
this first example here.

1666
01:16:38,540 --> 01:16:42,370
So acting as the computer, I take in
ls, which is this list I've highlighted.

1667
01:16:42,370 --> 01:16:45,730
And then I go ahead and
say, for i in range ls,

1668
01:16:45,730 --> 01:16:48,100
so that's going to start me off at 0.

1669
01:16:48,100 --> 01:16:50,860
And I'm going to say if
ls bracket 0, which is 3,

1670
01:16:50,860 --> 01:16:59,050
is less than ls bracket 0 minus
1, is going to be ls bracket

1671
01:16:59,050 --> 01:17:11,890
negative 1, which is this 2 here, which
I guess will not be a problem here.

1672
01:17:11,890 --> 01:17:18,250
Because the only case
when this first element is

1673
01:17:18,250 --> 01:17:22,420
going to be less than the last
element is when the array is not

1674
01:17:22,420 --> 01:17:23,500
rotated at all.

1675
01:17:23,500 --> 01:17:27,710
So I would want to
return 0 in that case.

1676
01:17:27,710 --> 01:17:33,790
So that actually makes me think that I
may not need this return 0 at the end

1677
01:17:33,790 --> 01:17:35,080
here.

1678
01:17:35,080 --> 01:17:39,130
Just kind of a peculiarity
of Python indexing,

1679
01:17:39,130 --> 01:17:42,670
where if I say ls
bracket negative 1, it's

1680
01:17:42,670 --> 01:17:46,720
going to be the last element of the
list, instead of giving me an error,

1681
01:17:46,720 --> 01:17:49,907
like it might in C or Java or
some other programming language.

1682
01:17:49,907 --> 01:17:51,240
TOMMY MACWILLIAMS: Yeah, thanks.

1683
01:17:51,240 --> 01:17:53,610
I hadn't thought of that,
but I think you're right.

1684
01:17:53,610 --> 01:17:55,560
CONNOR LEGGETT: All right.

1685
01:17:55,560 --> 01:17:58,810
And then finishing, go through
that example, for this one,

1686
01:17:58,810 --> 01:18:05,310
it would start out at 0 and see
that 3 is, in fact, greater than 2,

1687
01:18:05,310 --> 01:18:07,063
so it'll continue.

1688
01:18:07,063 --> 01:18:07,980
It'll get to this one.

1689
01:18:07,980 --> 01:18:11,020
It'll say that 4 is greater
than 3, so it will continue.

1690
01:18:11,020 --> 01:18:14,130
It'll get to 5, say 5 is greater
than 4, so it'll continue.

1691
01:18:14,130 --> 01:18:20,760
But then when i is 3, we'll see
that ls bracket i is, in fact,

1692
01:18:20,760 --> 01:18:22,830
less than ls bracket i minus 1.

1693
01:18:22,830 --> 01:18:28,330
So we'll go ahead and return i,
which is 3, which is as expected.

1694
01:18:28,330 --> 01:18:33,100
So just running through it as
a human, that seems to work.

1695
01:18:33,100 --> 01:18:35,520
So now let's move on
to writing some tests.

1696
01:18:35,520 --> 01:18:39,490
Yeah so I can go ahead and say--

1697
01:18:39,490 --> 01:18:43,140
I'll just print one out first to verify.

1698
01:18:43,140 --> 01:18:51,240
So I'll go ahead and say, print, get
rotate numbers of our first list here.

1699
01:18:51,240 --> 01:18:57,780
And hopefully, this will end up with 3,
so I can go ahead and run my code here.

1700
01:18:57,780 --> 01:19:03,750
And we've got a problem here, where
I'm saying list object cannot be

1701
01:19:03,750 --> 01:19:05,730
interpreted as an integer.

1702
01:19:05,730 --> 01:19:11,010
OK, and I think what's going on here is
I can't just say for i in range lists,

1703
01:19:11,010 --> 01:19:12,880
because range has to take in an integer.

1704
01:19:12,880 --> 01:19:15,742
So I actually have to look
at the length of the list--

1705
01:19:15,742 --> 01:19:17,325
TOMMY MACWILLIAMS: Yeah, that's right.

1706
01:19:17,325 --> 01:19:18,490
CONNOR LEGGETT: --instead.

1707
01:19:18,490 --> 01:19:20,830
So now I'll go ahead and try that again.

1708
01:19:20,830 --> 01:19:23,280
All right, nice, and so
that's looking pretty good.

1709
01:19:23,280 --> 01:19:25,440
So that's looking like it's printing 3.

1710
01:19:25,440 --> 01:19:29,920
And then I'll just go ahead and
try with our other lists here.

1711
01:19:29,920 --> 01:19:36,020
So hopefully, this second
one should translate to 4.

1712
01:19:36,020 --> 01:19:41,530
This third one should translate to 3.

1713
01:19:41,530 --> 01:19:46,530
And this last one should translate to 0.

1714
01:19:46,530 --> 01:19:49,270
And we'll go ahead and run that.

1715
01:19:49,270 --> 01:19:53,110
And it looks like I added--

1716
01:19:53,110 --> 01:19:53,777
oh, I forgot a--

1717
01:19:53,777 --> 01:19:55,652
TOMMY MACWILLIAMS: Yeah
it's the appropriate.

1718
01:19:55,652 --> 01:19:57,578
CONNOR LEGGETT: --bracket
here, I'll run that.

1719
01:19:57,578 --> 01:19:58,870
TOMMY MACWILLIAMS: Looks right.

1720
01:19:58,870 --> 01:20:02,690
CONNOR LEGGETT: And it looks like
that has given us the result we want.

1721
01:20:02,690 --> 01:20:04,370
So that's looking pretty good to me.

1722
01:20:04,370 --> 01:20:07,570
So I guess I'll check in at this
point and see if there's anything

1723
01:20:07,570 --> 01:20:10,300
that you notice about this
or any other corner cases

1724
01:20:10,300 --> 01:20:11,500
you think I should try out.

1725
01:20:11,500 --> 01:20:13,960
TOMMY MACWILLIAMS: Yeah,
this looks right to me.

1726
01:20:13,960 --> 01:20:16,107
What is the runtime of this?

1727
01:20:16,107 --> 01:20:17,440
CONNOR LEGGETT: Yeah, of course.

1728
01:20:17,440 --> 01:20:23,290
So going through the runtime, since I'm
going through i in range length of ls,

1729
01:20:23,290 --> 01:20:27,520
I'm going to go through this loop
a maximum of n times if there's

1730
01:20:27,520 --> 01:20:29,440
n elements in ls.

1731
01:20:29,440 --> 01:20:33,790
And then just accessing elements
of a list in Python is 0 of 1,

1732
01:20:33,790 --> 01:20:38,020
and comparing to integers in
Python is going to be 0 of 1,

1733
01:20:38,020 --> 01:20:40,520
in terms of the size of the list.

1734
01:20:40,520 --> 01:20:45,880
And so this whole operation inside the
for loop is going to be constant time.

1735
01:20:45,880 --> 01:20:50,530
So I going to say that the entire
runtime is going to be 0 of n.

1736
01:20:50,530 --> 01:20:54,820
So if there are n elements in the
list, it'll take a maximum of n times.

1737
01:20:54,820 --> 01:20:56,740
TOMMY MACWILLIAMS: Yeah,
I think that's right.

1738
01:20:56,740 --> 01:20:59,820
Can you think of any
ways to make it faster?

1739
01:20:59,820 --> 01:21:01,170
CONNOR LEGGETT: Yeah, let's see.

1740
01:21:01,170 --> 01:21:08,445
So if I'm working on making
it faster, I wonder--

1741
01:21:08,445 --> 01:21:14,120


1742
01:21:14,120 --> 01:21:26,900
so I guess any way that we could
make it faster would be to-- hmm.

1743
01:21:26,900 --> 01:21:32,930
I'm actually a little bit unsure,
because going through it as a human,

1744
01:21:32,930 --> 01:21:36,470
I can't think of anything else I
would do than go through each element.

1745
01:21:36,470 --> 01:21:40,400
So I'd love to know if you have
any hints, if you could nudge me

1746
01:21:40,400 --> 01:21:41,900
in the right direction a little bit.

1747
01:21:41,900 --> 01:21:46,040
TOMMY MACWILLIAMS: Yeah, so keep
in mind that the list is sorted.

1748
01:21:46,040 --> 01:21:49,530
Even though it's rotated, we have
this property that it's sorted.

1749
01:21:49,530 --> 01:21:53,660
So is there any search we could use that
leverages the fact that it's sorted?

1750
01:21:53,660 --> 01:21:57,010


1751
01:21:57,010 --> 01:21:58,970
CONNOR LEGGETT: We could use--

1752
01:21:58,970 --> 01:22:05,840
so generally, with the sorted list,
we could use just a binary search.

1753
01:22:05,840 --> 01:22:13,480
So if we're going
through and doing that,

1754
01:22:13,480 --> 01:22:19,250
then how would I apply a
binary search to this problem?

1755
01:22:19,250 --> 01:22:23,980
So I guess I could--

1756
01:22:23,980 --> 01:22:24,670
let's see.

1757
01:22:24,670 --> 01:22:28,480
So if I go through the motions
of doing a binary search

1758
01:22:28,480 --> 01:22:32,720
with a problem like this,
I would start in the middle

1759
01:22:32,720 --> 01:22:39,910
and see that I've got a size 5 here.

1760
01:22:39,910 --> 01:22:42,940
OK, so I guess I would have to
look at the first element first,

1761
01:22:42,940 --> 01:22:44,800
see that my first element is a 3.

1762
01:22:44,800 --> 01:22:47,590
And then when I head
to the middle, I see

1763
01:22:47,590 --> 01:22:55,060
that I'm at size 5, which means that
it's been increasing since then, which

1764
01:22:55,060 --> 01:22:58,990
means, I guess at that point, I can
go ahead and narrow my search down

1765
01:22:58,990 --> 01:23:04,140
to the latter half of the list
after I've checked that one.

1766
01:23:04,140 --> 01:23:08,520
And vice versa, if I, let's say,
for-- what's a good example?

1767
01:23:08,520 --> 01:23:13,380
For this one, if I start at--

1768
01:23:13,380 --> 01:23:16,860


1769
01:23:16,860 --> 01:23:19,290
I guess I don't have a
great example for that one.

1770
01:23:19,290 --> 01:23:24,360
But if I were to get to a point
where I was checking this 3 here,

1771
01:23:24,360 --> 01:23:27,060
then I would know if
I'm checking at this 3,

1772
01:23:27,060 --> 01:23:31,260
I know I can just look at
everything before the 3.

1773
01:23:31,260 --> 01:23:36,930
So yeah, I think a binary search
would be able to help us out there.

1774
01:23:36,930 --> 01:23:39,270
Just want to check in time-wise.

1775
01:23:39,270 --> 01:23:42,000
Is it all right if I go ahead
and try and implement that,

1776
01:23:42,000 --> 01:23:43,742
or do you think we should move on?

1777
01:23:43,742 --> 01:23:46,200
TOMMY MACWILLIAMS: Yeah, let's
try to implement it quickly,

1778
01:23:46,200 --> 01:23:48,445
and then if we run out of
time, that's totally fine

1779
01:23:48,445 --> 01:23:49,570
CONNOR LEGGETT: Yeah, sure.

1780
01:23:49,570 --> 01:23:57,590
So let me leave this other one here,
just to have something to reference.

1781
01:23:57,590 --> 01:24:06,170
And now this newer version here, I don't
necessarily want to go through each i

1782
01:24:06,170 --> 01:24:07,940
in this range.

1783
01:24:07,940 --> 01:24:09,770
I want to go ahead and say my--

1784
01:24:09,770 --> 01:24:14,180


1785
01:24:14,180 --> 01:24:21,000
I only care about when I'm going
from a high number to a lower number.

1786
01:24:21,000 --> 01:24:25,790
So I'll go ahead and say that my
highest number that I've seen so far

1787
01:24:25,790 --> 01:24:31,430
is going to be the first element
in the list, is ls bracket 0.

1788
01:24:31,430 --> 01:24:34,100


1789
01:24:34,100 --> 01:24:40,310
And then I can go ahead and say that
my lowest element that I've seen so far

1790
01:24:40,310 --> 01:24:42,950
is--

1791
01:24:42,950 --> 01:24:47,390
we can go ahead and look at
the last element of the list,

1792
01:24:47,390 --> 01:24:52,010
because I know it's going
to be somewhere in between.

1793
01:24:52,010 --> 01:24:55,220
Or I guess we don't necessarily
know that's the lowest.

1794
01:24:55,220 --> 01:25:00,440
Let me think about how we
would want to initialize those.

1795
01:25:00,440 --> 01:25:05,540
So 5, if that is the lowest, then--

1796
01:25:05,540 --> 01:25:11,360
yeah, I guess I'll just start out with
0 and go through the logic of this.

1797
01:25:11,360 --> 01:25:21,530
And then-- although I think
I'm going to actually want--

1798
01:25:21,530 --> 01:25:27,300
so if I'm looking for an
item that is lower here,

1799
01:25:27,300 --> 01:25:32,400
I'm also going to want to keep track
of the indices that I'm working with.

1800
01:25:32,400 --> 01:25:34,830
Does that seem like it's going
to be on the right track?

1801
01:25:34,830 --> 01:25:36,663
TOMMY MACWILLIAMS: Yeah,
I think in general,

1802
01:25:36,663 --> 01:25:38,260
you'll probably want to look at the--

1803
01:25:38,260 --> 01:25:40,772
keep track of the left and
right endpoints of the search.

1804
01:25:40,772 --> 01:25:43,230
And then you can compare wherever
you land on in the middle

1805
01:25:43,230 --> 01:25:45,140
to those left and right endpoints.

1806
01:25:45,140 --> 01:25:47,140
CONNOR LEGGETT: Yeah,
yeah, that's a good point.

1807
01:25:47,140 --> 01:25:54,900
So let me go ahead and say that left for
now is going to be 0, the first index.

1808
01:25:54,900 --> 01:25:57,570
And then since we're starting
our search with the full list,

1809
01:25:57,570 --> 01:26:02,685
our right is going to be the
length of the list minus 1.

1810
01:26:02,685 --> 01:26:05,340


1811
01:26:05,340 --> 01:26:10,650
And then what I can go ahead
and say, I can go ahead

1812
01:26:10,650 --> 01:26:13,740
and instead of having
this whole for loop here,

1813
01:26:13,740 --> 01:26:18,480
I can say, while right
is greater than left,

1814
01:26:18,480 --> 01:26:22,920
so while we're still
working on converging n.

1815
01:26:22,920 --> 01:26:27,450
Then I'll go ahead and say
that my current index is going

1816
01:26:27,450 --> 01:26:34,660
to be equal to right plus left, 5 by 2.

1817
01:26:34,660 --> 01:26:38,540
Let me get some parentheses in there.

1818
01:26:38,540 --> 01:26:41,862
And so that will take
the middle element.

1819
01:26:41,862 --> 01:26:43,570
And then since we're
working in Python, I

1820
01:26:43,570 --> 01:26:45,903
realize I should probably do
some integer division here.

1821
01:26:45,903 --> 01:26:47,740
TOMMY MACWILLIAMS: Yep.

1822
01:26:47,740 --> 01:26:51,410
CONNOR LEGGETT: So that'll get us our
current index that we're working with.

1823
01:26:51,410 --> 01:26:55,270
And then I can go ahead and say--

1824
01:26:55,270 --> 01:26:57,770
and let me look at this example
to see what we should do.

1825
01:26:57,770 --> 01:27:05,230
So in this example, if our current index
is greater than the index on the left,

1826
01:27:05,230 --> 01:27:08,060
then I know I can narrow my
search to the right half.

1827
01:27:08,060 --> 01:27:16,000
So if ls bracket current, so if
that current number is greater

1828
01:27:16,000 --> 01:27:22,150
than ls bracket left, it's greater than
the left number, then I can go ahead,

1829
01:27:22,150 --> 01:27:26,390
and I can narrow down the search
to this right half of the list.

1830
01:27:26,390 --> 01:27:33,970
So I can go ahead and say
that left equals current.

1831
01:27:33,970 --> 01:27:38,750
And then I can go ahead
and say, otherwise

1832
01:27:38,750 --> 01:27:51,460
if ls bracket current is
less than ls bracket left,

1833
01:27:51,460 --> 01:27:58,450
then I can go ahead and
say, right equals current.

1834
01:27:58,450 --> 01:28:05,290
So I know that we are going ahead
and changing our rightmost there.

1835
01:28:05,290 --> 01:28:10,180


1836
01:28:10,180 --> 01:28:13,300
And so now we just have
to work on what happens

1837
01:28:13,300 --> 01:28:16,340
after we do narrow in on that case.

1838
01:28:16,340 --> 01:28:19,090
So let's go through the example
with this top one again.

1839
01:28:19,090 --> 01:28:23,420
So we get to this
point with this 5 here,

1840
01:28:23,420 --> 01:28:27,400
and so now we know that our difference
is going to be between 5 and 2.

1841
01:28:27,400 --> 01:28:31,630
And actually, now that I'm thinking of--

1842
01:28:31,630 --> 01:28:34,820
yes, we're looking between 5 and 2 here.

1843
01:28:34,820 --> 01:28:41,350
So our left at this point is going to
be 2, and our right is going to be 4.

1844
01:28:41,350 --> 01:28:44,870
And so now our central
one is going to be 1 here.

1845
01:28:44,870 --> 01:28:51,410
And so now, since 1 is less
than 5, we would hit this case,

1846
01:28:51,410 --> 01:28:53,170
and so we would say
right equals current.

1847
01:28:53,170 --> 01:28:57,970
So now our left is going to be 5,
and our right is going to be 1.

1848
01:28:57,970 --> 01:29:05,560
So we know that we're ending
here when our right one is just 1

1849
01:29:05,560 --> 01:29:08,980
more than our left one.

1850
01:29:08,980 --> 01:29:13,120
And so let me go through what the
example would do in this if left is 2,

1851
01:29:13,120 --> 01:29:16,270
as we have here, and right is 3.

1852
01:29:16,270 --> 01:29:21,220
So when one left is 2 and right is
3, our current is going to be 3.

1853
01:29:21,220 --> 01:29:25,570
And so this is going to say, if current
is greater than the left one, which

1854
01:29:25,570 --> 01:29:28,690
isn't true, because they're the
same thing, and then we'll say,

1855
01:29:28,690 --> 01:29:34,090
if current is less than
the left one, which

1856
01:29:34,090 --> 01:29:37,340
isn't going to be true,
because they're the same,

1857
01:29:37,340 --> 01:29:39,920
then we can have an else statement here.

1858
01:29:39,920 --> 01:29:43,270
And if it's an else, then we
know that this current one

1859
01:29:43,270 --> 01:29:46,450
is kind of 1 below where we want to do.

1860
01:29:46,450 --> 01:29:55,770
So can we try and return our right here.

1861
01:29:55,770 --> 01:29:58,560
So now, just running
through with another example

1862
01:29:58,560 --> 01:30:04,090
as kind of a sanity check,
so we'll take this one here.

1863
01:30:04,090 --> 01:30:10,450
Let's say, so our left is
0 to start, right is 4.

1864
01:30:10,450 --> 01:30:16,210
We can find the one in the middle,
which is going to be element 2.

1865
01:30:16,210 --> 01:30:20,420
And then we're going to see
2 is greater than our left.

1866
01:30:20,420 --> 01:30:26,350
So that means that we're going to narrow
our search down to these last three

1867
01:30:26,350 --> 01:30:29,030
elements here.

1868
01:30:29,030 --> 01:30:33,910
And now we're going to
go ahead and say that--

1869
01:30:33,910 --> 01:30:37,820
yeah, so now we're going to go ahead
and find the middle element here,

1870
01:30:37,820 --> 01:30:38,590
which is 1.

1871
01:30:38,590 --> 01:30:44,290
And we'll say that left--

1872
01:30:44,290 --> 01:30:46,600
yeah, so this is our current,
and then we're checking

1873
01:30:46,600 --> 01:30:48,880
is current greater than left.

1874
01:30:48,880 --> 01:30:49,990
No, that's not true.

1875
01:30:49,990 --> 01:30:51,370
Is current less than left?

1876
01:30:51,370 --> 01:30:55,750
Yes, that is true, so we're
changing our current to this,

1877
01:30:55,750 --> 01:30:58,120
or we're changing our right to this.

1878
01:30:58,120 --> 01:31:01,000
And so now we're just
looking at this section here.

1879
01:31:01,000 --> 01:31:02,650
These are our left and right.

1880
01:31:02,650 --> 01:31:07,970
And when we compare left to
current and right to current,

1881
01:31:07,970 --> 01:31:10,720
they're both going to be the same
thing, so we get this else here.

1882
01:31:10,720 --> 01:31:15,410
And we return what's on the right,
which is 1, which should be correct.

1883
01:31:15,410 --> 01:31:26,290
So just basic conceptually wise, it
looks like things are working out here.

1884
01:31:26,290 --> 01:31:29,950
Oh, let me run through it real
quick with the unrotated around

1885
01:31:29,950 --> 01:31:31,970
to see how things are going there.

1886
01:31:31,970 --> 01:31:34,900
So in this case, we're going
to get to this middle one.

1887
01:31:34,900 --> 01:31:39,710
We're going to see that the
current is greater than the left.

1888
01:31:39,710 --> 01:31:42,220
So we'll get to this
one here now, and we'll

1889
01:31:42,220 --> 01:31:47,120
see that the right is still
greater than the left.

1890
01:31:47,120 --> 01:31:53,660
So now our current is still
greater than the left,

1891
01:31:53,660 --> 01:31:57,010
so our left is becoming this one.

1892
01:31:57,010 --> 01:32:00,490
And now we're going
to go ahead and say--

1893
01:32:00,490 --> 01:32:04,210
the left is still going to be this one.

1894
01:32:04,210 --> 01:32:07,103
So that's not going to give us the
right answer here, it looks like.

1895
01:32:07,103 --> 01:32:09,520
TOMMY MACWILLIAMS: Yeah, I
think the stop condition you're

1896
01:32:09,520 --> 01:32:11,770
looking for is in that first example.

1897
01:32:11,770 --> 01:32:14,770
You can see with the one, the
number to the left is bigger,

1898
01:32:14,770 --> 01:32:16,550
and the number to the right is bigger.

1899
01:32:16,550 --> 01:32:18,880
So that's kind of the stop
condition for your search.

1900
01:32:18,880 --> 01:32:22,088
Otherwise, though, I think
the binary search is right.

1901
01:32:22,088 --> 01:32:22,880
CONNOR LEGGETT: OK.

1902
01:32:22,880 --> 01:32:24,400
Yeah, so that makes sense.

1903
01:32:24,400 --> 01:32:30,060
So maybe I can go ahead
and check that condition.

1904
01:32:30,060 --> 01:32:37,910
So if we're looking at our current
one, do you think for each time,

1905
01:32:37,910 --> 01:32:45,320
we would look at our current and check
if it's smaller than the left one,

1906
01:32:45,320 --> 01:32:47,267
and smaller than the right one?

1907
01:32:47,267 --> 01:32:49,100
TOMMY MACWILLIAMS: Yeah,
that's seems right.

1908
01:32:49,100 --> 01:32:50,780
CONNOR LEGGETT: OK, sure.

1909
01:32:50,780 --> 01:32:58,410
So we're already checking if our current
is smaller than the left one here.

1910
01:32:58,410 --> 01:33:02,960
And so I can go ahead and
nest a little bit of ifs

1911
01:33:02,960 --> 01:33:17,980
here and say, if our current
is also smaller than the one--

1912
01:33:17,980 --> 01:33:18,940
oh, sorry.

1913
01:33:18,940 --> 01:33:21,800
But we're not looking at
direct neighbors there.

1914
01:33:21,800 --> 01:33:23,960
So let's look at direct
neighbors up here.

1915
01:33:23,960 --> 01:33:32,470
So that would look like if
ls bracket current minus 1

1916
01:33:32,470 --> 01:33:39,060
is less than ls bracket
current, which is

1917
01:33:39,060 --> 01:33:43,410
less than ls bracket
current plus 1-- which

1918
01:33:43,410 --> 01:33:47,130
is a nice thing we can do
in Python is just combine

1919
01:33:47,130 --> 01:33:51,090
these different inequalities.

1920
01:33:51,090 --> 01:33:53,800
And I'll take away those
parentheses for now.

1921
01:33:53,800 --> 01:33:55,860
And so in this case,
we know that current

1922
01:33:55,860 --> 01:33:57,490
is the one that we want to return.

1923
01:33:57,490 --> 01:34:00,360
So we can go ahead and
return current here.

1924
01:34:00,360 --> 01:34:03,100


1925
01:34:03,100 --> 01:34:10,950
And then let me just do another
sanity check with these--

1926
01:34:10,950 --> 01:34:13,450
TOMMY MACWILLIAMS: Yeah, I think
this looks basically right.

1927
01:34:13,450 --> 01:34:17,470
Just for time, we're basically at
time, but yeah, overall, this search

1928
01:34:17,470 --> 01:34:19,357
looks pretty good.

1929
01:34:19,357 --> 01:34:20,440
CONNOR LEGGETT: All right.

1930
01:34:20,440 --> 01:34:20,910
TOMMY MACWILLIAMS: Awesome.

1931
01:34:20,910 --> 01:34:23,410
Well, thanks for taking the
time for chatting with me today,

1932
01:34:23,410 --> 01:34:24,970
and have a great rest of your day.

1933
01:34:24,970 --> 01:34:26,560
CONNOR LEGGETT: Yeah, thanks,
it was great to meet you.

1934
01:34:26,560 --> 01:34:28,610
TOMMY MACWILLIAMS:
Awesome, see you later.

1935
01:34:28,610 --> 01:34:30,390
CONNOR LEGGETT: Bye.

1936
01:34:30,390 --> 01:34:35,850
And then for feedback, is it
best to keep my screen shared?

1937
01:34:35,850 --> 01:34:39,400
TOMMY MACWILLIAMS: Let's actually just
switch to Maria, just for time's sake.

1938
01:34:39,400 --> 01:34:43,390
So Maria can flip into screen sharing,
and we can go into the next question.

1939
01:34:43,390 --> 01:34:46,140
All right, so I'll describe
the problem, and then you

1940
01:34:46,140 --> 01:34:47,560
can take it away from there.

1941
01:34:47,560 --> 01:34:50,200
So what we're going to
do is, given a string,

1942
01:34:50,200 --> 01:34:54,360
we're going to be looking for the
length of the longest substring

1943
01:34:54,360 --> 01:34:56,590
without repeating characters.

1944
01:34:56,590 --> 01:35:03,330
So, for instance, if I gave you
the string abcabcbb, the length

1945
01:35:03,330 --> 01:35:05,700
of the longest substring there is 3.

1946
01:35:05,700 --> 01:35:08,730
So there's abc where it doesn't have
any repeating characters in there,

1947
01:35:08,730 --> 01:35:10,438
and there are a few
other ones like that.

1948
01:35:10,438 --> 01:35:12,188
So your function would
take that as input,

1949
01:35:12,188 --> 01:35:14,790
and then just return the length
of that string, which is 3.

1950
01:35:14,790 --> 01:35:16,190
MARIA ZLATKOVA: Awesome.

1951
01:35:16,190 --> 01:35:17,190
Hey, Tommy, how are you?

1952
01:35:17,190 --> 01:35:17,460
TOMMY MACWILLIAMS: Awesome.

1953
01:35:17,460 --> 01:35:18,255
How's it going?

1954
01:35:18,255 --> 01:35:19,130
MARIA ZLATKOVA: Good.

1955
01:35:19,130 --> 01:35:20,193
How's your day going?

1956
01:35:20,193 --> 01:35:21,360
TOMMY MACWILLIAMS: I'm good.

1957
01:35:21,360 --> 01:35:24,957
It's a sunny day still in San
Francisco, so I'm loving it.

1958
01:35:24,957 --> 01:35:26,040
MARIA ZLATKOVA: Beautiful.

1959
01:35:26,040 --> 01:35:29,152
It's kind of a gloomy day in
New York, but still pretty warm,

1960
01:35:29,152 --> 01:35:30,570
which is about all we get.

1961
01:35:30,570 --> 01:35:31,200
TOMMY MACWILLIAMS: Nice.

1962
01:35:31,200 --> 01:35:33,075
Well, thanks so much
for jumping on the call.

1963
01:35:33,075 --> 01:35:35,570
Let's jump into a coding problem.

1964
01:35:35,570 --> 01:35:36,570
MARIA ZLATKOVA: Awesome.

1965
01:35:36,570 --> 01:35:37,560
TOMMY MACWILLIAMS:
All right, so it looks

1966
01:35:37,560 --> 01:35:38,977
like you've got the Replit set up.

1967
01:35:38,977 --> 01:35:40,230
I notice you're set to Python.

1968
01:35:40,230 --> 01:35:42,240
Is that the language you'd like to use?

1969
01:35:42,240 --> 01:35:43,770
MARIA ZLATKOVA: Yeah, that's
the language I'd like to use,

1970
01:35:43,770 --> 01:35:44,760
if that sounds good with you.

1971
01:35:44,760 --> 01:35:46,760
TOMMY MACWILLIAMS: Perfect,
that's good with me.

1972
01:35:46,760 --> 01:35:47,828
MARIA ZLATKOVA: Cool.

1973
01:35:47,828 --> 01:35:49,870
Just thinking through this
question a little bit,

1974
01:35:49,870 --> 01:35:52,320
I guess a few clarifying questions.

1975
01:35:52,320 --> 01:35:53,853
TOMMY MACWILLIAMS: Yeah.

1976
01:35:53,853 --> 01:35:56,270
MARIA ZLATKOVA: So we just
want to find the length, right?

1977
01:35:56,270 --> 01:35:59,642
We don't care about returning
one or all of the sequences

1978
01:35:59,642 --> 01:36:00,600
that match that length?

1979
01:36:00,600 --> 01:36:02,142
TOMMY MACWILLIAMS: Yep, that's right.

1980
01:36:02,142 --> 01:36:04,140
MARIA ZLATKOVA: Awesome, cool.

1981
01:36:04,140 --> 01:36:06,210
And then when it comes
to the input, I see

1982
01:36:06,210 --> 01:36:09,180
the example we have here are
all lowercase characters.

1983
01:36:09,180 --> 01:36:12,690
Should we assume that we're going to
get all lowercase characters as input,

1984
01:36:12,690 --> 01:36:15,148
or do we want to handle kind
of like sanitizing that input?

1985
01:36:15,148 --> 01:36:17,148
TOMMY MACWILLIAMS: Yeah,
that's a good question.

1986
01:36:17,148 --> 01:36:18,870
You can assume the input's all lowercase

1987
01:36:18,870 --> 01:36:21,000
MARIA ZLATKOVA: Awesome, cool.

1988
01:36:21,000 --> 01:36:26,687
And I guess, could the input be
empty, like the empty script string?

1989
01:36:26,687 --> 01:36:29,020
TOMMY MACWILLIAMS: It could,
in which case you return 0.

1990
01:36:29,020 --> 01:36:30,270
MARIA ZLATKOVA: Awesome, cool.

1991
01:36:30,270 --> 01:36:32,330


1992
01:36:32,330 --> 01:36:34,730
Great, I think that covered
all of my questions.

1993
01:36:34,730 --> 01:36:38,840
So maybe then just give a
few examples, so abc, abcdb,

1994
01:36:38,840 --> 01:36:43,130
the answer would be 3, the
output that we want to give.

1995
01:36:43,130 --> 01:36:48,638
If we have all repeating characters the
output would be just 1, I'm assuming?

1996
01:36:48,638 --> 01:36:50,180
TOMMY MACWILLIAMS: Yep, that's right.

1997
01:36:50,180 --> 01:36:51,470
MARIA ZLATKOVA: Awesome.

1998
01:36:51,470 --> 01:36:54,680
And conversely, if we have
all different characters,

1999
01:36:54,680 --> 01:36:57,590
the output would be 6 in this case?

2000
01:36:57,590 --> 01:36:59,062
TOMMY MACWILLIAMS: Yep.

2001
01:36:59,062 --> 01:37:01,520
MARIA ZLATKOVA: And then, as
we said, if we get the string,

2002
01:37:01,520 --> 01:37:04,334
the output would be 0.

2003
01:37:04,334 --> 01:37:10,730
Yeah, awesome, I think
that makes a lot of sense.

2004
01:37:10,730 --> 01:37:14,460
When I'm looking at this
initial string and problem,

2005
01:37:14,460 --> 01:37:17,810
I guess the first instinct that I
have for the very naive solution

2006
01:37:17,810 --> 01:37:21,530
is that brute force.

2007
01:37:21,530 --> 01:37:26,900
We can just get all of the
possible subsequences or substrings

2008
01:37:26,900 --> 01:37:29,600
and check whether there are
any duplicate characters

2009
01:37:29,600 --> 01:37:30,800
within those substrings.

2010
01:37:30,800 --> 01:37:36,740
So in the abc, abcbb example,
that would be first starting

2011
01:37:36,740 --> 01:37:39,650
with every substring
starting with a would be a.

2012
01:37:39,650 --> 01:37:43,582
And checking if that has any
repeating characters, then ab,

2013
01:37:43,582 --> 01:37:45,290
it doesn't have any
duplicate characters.

2014
01:37:45,290 --> 01:37:49,040
Then abc, it also doesn't have any
duplicate characters, then abca,

2015
01:37:49,040 --> 01:37:51,860
which has duplicate characters.

2016
01:37:51,860 --> 01:37:54,710
And then every single
one-- abcab would also

2017
01:37:54,710 --> 01:37:58,400
have duplicate characters, given
that the substring of that had them,

2018
01:37:58,400 --> 01:38:05,030
and also that b is repeating similar,
with abcabc abcabcb and abcabcdb.

2019
01:38:05,030 --> 01:38:10,610
So it would be a very similar
case if we started b and at c.

2020
01:38:10,610 --> 01:38:14,300
So it's kind of making
me feel like the naive

2021
01:38:14,300 --> 01:38:19,200
solution would be doing a little
bit of work that is not useful.

2022
01:38:19,200 --> 01:38:22,580
So if we could avoid
that, that would be great.

2023
01:38:22,580 --> 01:38:26,660
Reasoning a little bit over that
nice, ancient, brute force approach,

2024
01:38:26,660 --> 01:38:31,590
going through all possible
subsequences would be quadratic time,

2025
01:38:31,590 --> 01:38:36,200
and then within that, checking
whether there are duplicate characters

2026
01:38:36,200 --> 01:38:38,790
within each substring
would be linear time.

2027
01:38:38,790 --> 01:38:43,760
So in total, that probably
would be n to the third.

2028
01:38:43,760 --> 01:38:45,260
So that is not great.

2029
01:38:45,260 --> 01:38:48,770
So I'm going to try to
optimize that, if we can.

2030
01:38:48,770 --> 01:38:53,990
And looking at this the
way I like described

2031
01:38:53,990 --> 01:38:57,050
it, starting at different
points of the string,

2032
01:38:57,050 --> 01:39:01,040
I'm starting to think maybe we
can have a range within the string

2033
01:39:01,040 --> 01:39:05,060
that we track with a starting
index and an ending index.

2034
01:39:05,060 --> 01:39:11,090
And maybe we can expand
the start of string

2035
01:39:11,090 --> 01:39:13,310
and expand our ending
index to the right.

2036
01:39:13,310 --> 01:39:18,680
And the moment we kind of encounter
a duplicate character or repeating

2037
01:39:18,680 --> 01:39:23,750
character, we start to move
our start index forward

2038
01:39:23,750 --> 01:39:27,140
to try to get to a subsequence
that doesn't contain

2039
01:39:27,140 --> 01:39:30,590
that duplicate character anymore, and
all the while, while we're doing that,

2040
01:39:30,590 --> 01:39:35,240
keep track of the longest
encountered length.

2041
01:39:35,240 --> 01:39:38,850
And I guess I have another
kind of assumption question is,

2042
01:39:38,850 --> 01:39:41,147
could we use an extra data
structure in this question?

2043
01:39:41,147 --> 01:39:42,980
TOMMY MACWILLIAMS: Yeah,
that's really fine.

2044
01:39:42,980 --> 01:39:45,230
I think that sort of sliding
window you're describing,

2045
01:39:45,230 --> 01:39:47,480
that sounds like a pretty
good approach to me.

2046
01:39:47,480 --> 01:39:48,480
MARIA ZLATKOVA: Awesome.

2047
01:39:48,480 --> 01:39:52,340
Yeah, I think for that sliding
window I'd like to use HashSet,

2048
01:39:52,340 --> 01:39:57,890
so that we can have pretty
constant time in looking up

2049
01:39:57,890 --> 01:40:02,270
how often certain characters
are within the string already,

2050
01:40:02,270 --> 01:40:04,818
and we can do that look-up very quickly.

2051
01:40:04,818 --> 01:40:06,610
TOMMY MACWILLIAMS:
Yeah, that sounds great.

2052
01:40:06,610 --> 01:40:07,610
MARIA ZLATKOVA: Awesome.

2053
01:40:07,610 --> 01:40:11,100
So if that sounds good, I'm going
to go ahead and define our function.

2054
01:40:11,100 --> 01:40:16,130
I'm going to call it length
of longest substring.

2055
01:40:16,130 --> 01:40:19,490
We take in just one parameter,
I'm assuming, which is the string.

2056
01:40:19,490 --> 01:40:20,630
TOMMY MACWILLIAMS: Yep.

2057
01:40:20,630 --> 01:40:26,370
MARIA ZLATKOVA: Then we need to define
the output, which is an integer.

2058
01:40:26,370 --> 01:40:28,670
So as I said, would
love to use a HashSet.

2059
01:40:28,670 --> 01:40:32,420
So I'm just going to
start by initializing it

2060
01:40:32,420 --> 01:40:36,110
and just initialize it
all to zeros for now.

2061
01:40:36,110 --> 01:40:40,610
And we're going to count how many
occurrences of each character

2062
01:40:40,610 --> 01:40:43,910
that we encounter happening
within the sub string.

2063
01:40:43,910 --> 01:40:47,810
And I'm going to multiply
that by 128, because I

2064
01:40:47,810 --> 01:40:52,740
know that we have 128 ASCII characters.

2065
01:40:52,740 --> 01:40:56,930
So just to start off,
initialize it with that.

2066
01:40:56,930 --> 01:41:00,560
And then going to
define the length, which

2067
01:41:00,560 --> 01:41:04,250
we're going to want to talk about
longest substrings and define it 0.

2068
01:41:04,250 --> 01:41:08,810
I'm getting some-- oh, OK,
variable assigned, but never used.

2069
01:41:08,810 --> 01:41:11,480
That's OK, because we
haven't used it yet.

2070
01:41:11,480 --> 01:41:16,310
And then I'm going to use the
start and end index approach that I

2071
01:41:16,310 --> 01:41:18,010
described with that sliding window.

2072
01:41:18,010 --> 01:41:22,950
So the start, maybe I'll just
call it left for clarity.

2073
01:41:22,950 --> 01:41:24,260
I'm going to assign that 0.

2074
01:41:24,260 --> 01:41:28,025
And then what we want to do first
is start going through the string.

2075
01:41:28,025 --> 01:41:33,000
You can do it in a few ways, while loop,
but maybe I'll do a for loop to start.

2076
01:41:33,000 --> 01:41:37,790
So you can say for right,
and right is referring

2077
01:41:37,790 --> 01:41:41,570
to my n index, which I
described for right in range

2078
01:41:41,570 --> 01:41:46,133
of the length of the string.

2079
01:41:46,133 --> 01:41:48,050
Thankfully, I saw Connor's
previous interview,

2080
01:41:48,050 --> 01:41:50,570
so I now know that we can work here.

2081
01:41:50,570 --> 01:41:54,560
But we have to do the
length of the string.

2082
01:41:54,560 --> 01:42:01,550
And then the first thing I want to do
in this case is to make sure that--

2083
01:42:01,550 --> 01:42:05,360
basically, what we
want to do is increment

2084
01:42:05,360 --> 01:42:10,340
the count of the specific character
that we're at by 1, so in the HashSet

2085
01:42:10,340 --> 01:42:11,340
that we have.

2086
01:42:11,340 --> 01:42:13,880
So the way we're going to do
that is we're going to say,

2087
01:42:13,880 --> 01:42:20,690
chars at s at the next right,
because S's are the string.

2088
01:42:20,690 --> 01:42:23,960
And we're going to increment that by 1.

2089
01:42:23,960 --> 01:42:30,230
And that's OK, because we've defined
all the characters, all of them,

2090
01:42:30,230 --> 01:42:33,930
essentially, fields in
our HashSet to be 0.

2091
01:42:33,930 --> 01:42:34,430
Cool.

2092
01:42:34,430 --> 01:42:38,272
So then the next thing
I want to do is I want

2093
01:42:38,272 --> 01:42:40,730
to take care of that case,
which I mentioned earlier, which

2094
01:42:40,730 --> 01:42:45,620
is, if we encounter a duplicate
character moving the left or the char

2095
01:42:45,620 --> 01:42:51,050
index forward until we don't
[INAUDIBLE] the character anymore.

2096
01:42:51,050 --> 01:42:55,100
And since we're using
chars as our HashSet,

2097
01:42:55,100 --> 01:42:58,370
the way that we know that we've
encountered more than one character

2098
01:42:58,370 --> 01:43:08,390
is if chars at index
right is greater than 1.

2099
01:43:08,390 --> 01:43:10,930
And so I'm going to
do a while loop here,

2100
01:43:10,930 --> 01:43:15,760
because we could have multiple repeating
characters within the substring.

2101
01:43:15,760 --> 01:43:20,950
So I want to make sure that I get rid
of all potential repeating characters.

2102
01:43:20,950 --> 01:43:23,050
Or the repeating
character could actually

2103
01:43:23,050 --> 01:43:26,350
be at a later point in the substring,
so I want to get to that point

2104
01:43:26,350 --> 01:43:31,510
and make sure I move the char
index by enough characters forward.

2105
01:43:31,510 --> 01:43:36,370
So the way we're going
to do this is first,

2106
01:43:36,370 --> 01:43:39,640
we want to update the
count that we're keeping

2107
01:43:39,640 --> 01:43:42,420
track of in our HashSet for
that specific character.

2108
01:43:42,420 --> 01:43:44,170
So the way we're going
to do that is we're

2109
01:43:44,170 --> 01:43:47,920
going to stay chars at s at left--

2110
01:43:47,920 --> 01:43:50,290
I think we called it left--

2111
01:43:50,290 --> 01:43:52,660
[INAUDIBLE] minus equals 1.

2112
01:43:52,660 --> 01:43:55,900
And then what I'm going
to do to move left forward

2113
01:43:55,900 --> 01:44:00,190
is I'm just going to increment left.

2114
01:44:00,190 --> 01:44:03,820
Cool, and then one thing
which I haven't done yet

2115
01:44:03,820 --> 01:44:08,480
is I want to make sure I'm keeping
track of the length, as we said earlier.

2116
01:44:08,480 --> 01:44:14,550
So the length at the
first iteration will be--

2117
01:44:14,550 --> 01:44:18,360
we've initialized it to
0, and then we calculated

2118
01:44:18,360 --> 01:44:20,220
the length between two
indexes in a string

2119
01:44:20,220 --> 01:44:24,660
typically would be the ending index
that is the starting index plus 1,

2120
01:44:24,660 --> 01:44:26,470
because it [INAUDIBLE] 0 index.

2121
01:44:26,470 --> 01:44:30,460
So in our case, it's
right minus left plus 1.

2122
01:44:30,460 --> 01:44:32,680
If I could spell correctly
that would be great.

2123
01:44:32,680 --> 01:44:35,700
And then what I'm going
to do, so that we always

2124
01:44:35,700 --> 01:44:39,270
keep track of the biggest
the maximum possible link,

2125
01:44:39,270 --> 01:44:42,040
is I'm going to use the
max function and say, I

2126
01:44:42,040 --> 01:44:47,940
want to take the max of the length
that we're keeping track of,

2127
01:44:47,940 --> 01:44:51,150
or the current link based on
the right and left integers.

2128
01:44:51,150 --> 01:44:54,000
And then I think the
last thing we want to do

2129
01:44:54,000 --> 01:45:01,500
is return the link, because we
want to retain that integer.

2130
01:45:01,500 --> 01:45:08,255
Cool, and I guess just a sanity
check a little bit before I run this,

2131
01:45:08,255 --> 01:45:11,370
I'm just going to do it with abca,
because it's a little bit shorter

2132
01:45:11,370 --> 01:45:13,650
than abcabcdb.

2133
01:45:13,650 --> 01:45:17,680
So abca should be 3,
the output for that.

2134
01:45:17,680 --> 01:45:23,350
So what we're going to do is,
when we start our for loop,

2135
01:45:23,350 --> 01:45:28,630
we're going to say that
chars at a is equal to 1.

2136
01:45:28,630 --> 01:45:30,400
We're going to add 1 to that.

2137
01:45:30,400 --> 01:45:32,960
It's not greater than 1
in the first iteration.

2138
01:45:32,960 --> 01:45:42,150
So what we're going to do is
just take the length and set it.

2139
01:45:42,150 --> 01:45:46,530
And because right and left are 0 at
that first run through the for loop,

2140
01:45:46,530 --> 01:45:50,010
it's going to be 0 minus 0 plus 1, so
length is going to be 1 at this stage.

2141
01:45:50,010 --> 01:45:54,840
And then when we get to b, chars
at b is going to be granted to 1.

2142
01:45:54,840 --> 01:45:57,972
Chars b still is not
greater than 1, so again,

2143
01:45:57,972 --> 01:46:00,180
we're just going to jump to
length, and at this point

2144
01:46:00,180 --> 01:46:05,190
it's going to be 2, because
1 minus 0 plus 1 is 2.

2145
01:46:05,190 --> 01:46:08,490
Then we're at index 2, which is
the third character, which is c.

2146
01:46:08,490 --> 01:46:14,370
So again, setting chars at c can be
1, and then it's not greater than 1,

2147
01:46:14,370 --> 01:46:17,580
so we just jump and then increment
the length again, to 3 this time,

2148
01:46:17,580 --> 01:46:21,520
because 2 minus 0 plus 1 is 3.

2149
01:46:21,520 --> 01:46:23,610
Then in the final
run-through of that for loop,

2150
01:46:23,610 --> 01:46:25,080
we're going to encounter a again.

2151
01:46:25,080 --> 01:46:29,730
And so chars at a is now equal
to 2, which is greater than 1.

2152
01:46:29,730 --> 01:46:34,350
So what we're going to do is decrement
that, so again, chars a becomes 1.

2153
01:46:34,350 --> 01:46:45,000
And we're going to move our left
index forward, and what we get now

2154
01:46:45,000 --> 01:46:53,940
is that we want to take the max of
either 3 or 0, 1, 2, 3, or 1, 2,

2155
01:46:53,940 --> 01:46:56,670
3, which, again, is going
to give us a length 3.

2156
01:46:56,670 --> 01:46:59,900
And we've run through basically
all of the elements here,

2157
01:46:59,900 --> 01:47:01,900
so what we're going to
do is just return line 3.

2158
01:47:01,900 --> 01:47:05,220
So just by running through the simple
example, I think this should work.

2159
01:47:05,220 --> 01:47:08,140
And I'm going to write
out some test pieces.

2160
01:47:08,140 --> 01:47:15,600
So maybe I'll just print, input,
and maybe just define our string

2161
01:47:15,600 --> 01:47:20,040
here first to be abcabcbb.

2162
01:47:20,040 --> 01:47:24,660
Our input is going to
be string, and then we

2163
01:47:24,660 --> 01:47:32,300
want to say length is equal to
calling the function on string,

2164
01:47:32,300 --> 01:47:34,600
and finally print.

2165
01:47:34,600 --> 01:47:41,620
Output to be length.

2166
01:47:41,620 --> 01:47:44,310
Cool, so it may have some errors.

2167
01:47:44,310 --> 01:47:47,740
I'm just going to run it through
the first time and see what happens.

2168
01:47:47,740 --> 01:47:52,760
And I see here that on line
14, we have, list indices

2169
01:47:52,760 --> 01:47:55,670
must be integers or slices, not strings.

2170
01:47:55,670 --> 01:47:58,623
So looking at line 14--

2171
01:47:58,623 --> 01:48:01,790
oh, I think this is very interesting,
because I referred to ASCII characters

2172
01:48:01,790 --> 01:48:05,270
in the beginning, but
then I didn't really

2173
01:48:05,270 --> 01:48:10,730
take into account that we can't have
a character itself as the index,

2174
01:48:10,730 --> 01:48:14,457
but I think we have to
use a function called ord.

2175
01:48:14,457 --> 01:48:16,040
TOMMY MACWILLIAMS: Yeah, that's right.

2176
01:48:16,040 --> 01:48:16,520
MARIA ZLATKOVA: Awesome.

2177
01:48:16,520 --> 01:48:18,920
Yeah, I think that function
returns an integer that

2178
01:48:18,920 --> 01:48:20,580
represents Unicode character.

2179
01:48:20,580 --> 01:48:27,560
So what I want to do is I want to
wrap ord around every single time I'm

2180
01:48:27,560 --> 01:48:29,105
using an index within chars.

2181
01:48:29,105 --> 01:48:31,760


2182
01:48:31,760 --> 01:48:35,950
Those should be all the times.

2183
01:48:35,950 --> 01:48:39,480
And then input abcabcbb,
output 3, that's great.

2184
01:48:39,480 --> 01:48:45,700
I'm just going to run through and add
a few more test cases like the ones we

2185
01:48:45,700 --> 01:48:49,830
mentioned, so bbb , abcdef--

2186
01:48:49,830 --> 01:48:54,120


2187
01:48:54,120 --> 01:49:00,030
no-- and maybe just an empty
string for good measure.

2188
01:49:00,030 --> 01:49:05,580
We'll see if that gets us
the right result. Cool.

2189
01:49:05,580 --> 01:49:13,080
So we get input bbb output 1, which
is great, and put abcdef output 6,

2190
01:49:13,080 --> 01:49:15,682
which is great.

2191
01:49:15,682 --> 01:49:19,562
pwekf-- I think this is right.

2192
01:49:19,562 --> 01:49:21,270
TOMMY MACWILLIAMS:
Yep, that looks right.

2193
01:49:21,270 --> 01:49:25,500
MARIA ZLATKOVA: OK, and then
input empty string, output 0.

2194
01:49:25,500 --> 01:49:26,680
Awesome.

2195
01:49:26,680 --> 01:49:27,930
I think this looks good to me.

2196
01:49:27,930 --> 01:49:30,758
Are there any other edge cases
that you want us to take a look at?

2197
01:49:30,758 --> 01:49:32,550
MARIA ZLATKOVA: No,
this looks great to me.

2198
01:49:32,550 --> 01:49:34,425
And I think you might
have mentioned already,

2199
01:49:34,425 --> 01:49:36,730
but what's the runtime
of this final solution?

2200
01:49:36,730 --> 01:49:42,010
MARIA ZLATKOVA: Yes, so space complexity
is linear if you're using HashSet.

2201
01:49:42,010 --> 01:49:45,160
And then time complexity--

2202
01:49:45,160 --> 01:49:48,880
so we have one for loop,
which is linear time,

2203
01:49:48,880 --> 01:49:55,010
and then searching within the HashSet
itself should be constant time.

2204
01:49:55,010 --> 01:49:59,385
So I think it comes down to linear time.

2205
01:49:59,385 --> 01:50:02,050
At most, you have to go through
each of the characters twice.

2206
01:50:02,050 --> 01:50:04,137
TOMMY MACWILLIAMS: Yeah,
that sounds right to me.

2207
01:50:04,137 --> 01:50:04,720
Well, awesome.

2208
01:50:04,720 --> 01:50:05,830
This looks great.

2209
01:50:05,830 --> 01:50:08,740
Well, thanks so much for taking
the time to chat with me today.

2210
01:50:08,740 --> 01:50:09,160
MARIA ZLATKOVA: Of course.

2211
01:50:09,160 --> 01:50:10,018
Thanks so much too.

2212
01:50:10,018 --> 01:50:11,560
It was a fun problem to work through.

2213
01:50:11,560 --> 01:50:12,685
TOMMY MACWILLIAMS: Awesome.

2214
01:50:12,685 --> 01:50:13,400
Have a good one.

2215
01:50:13,400 --> 01:50:16,030
All right, so those were
the two mock interviews.

2216
01:50:16,030 --> 01:50:18,765
I'll just say a few closing
words for after the interview.

2217
01:50:18,765 --> 01:50:20,890
And then I know we're going
a little bit over time,

2218
01:50:20,890 --> 01:50:24,400
but I'll be around if anyone
wants to ask questions.

2219
01:50:24,400 --> 01:50:26,830
And then maybe before
that we can quickly

2220
01:50:26,830 --> 01:50:28,660
go through just some
things that went well

2221
01:50:28,660 --> 01:50:31,202
in the mock interviews and some
things that didn't, so people

2222
01:50:31,202 --> 01:50:33,580
could give their own
feedback and thoughts.

2223
01:50:33,580 --> 01:50:38,450
But before that, I'll just
have a few closing thoughts.

2224
01:50:38,450 --> 01:50:41,290
So one is, we didn't simulate
this in the mock interview,

2225
01:50:41,290 --> 01:50:45,110
but after the interview is over is
kind of your chance to ask questions.

2226
01:50:45,110 --> 01:50:48,610
So now you would ask
questions of your interviewer,

2227
01:50:48,610 --> 01:50:51,290
and here are some good
questions that I like to ask--

2228
01:50:51,290 --> 01:50:54,760
asking about their team's culture or
what their typical day-to-day looks

2229
01:50:54,760 --> 01:50:58,300
like, things that they've worked
on, things that they're proud of.

2230
01:50:58,300 --> 01:51:00,040
And this is a really
good way just to get

2231
01:51:00,040 --> 01:51:03,040
a sense of what it's really
like to work at that company

2232
01:51:03,040 --> 01:51:04,570
you're interviewing with.

2233
01:51:04,570 --> 01:51:08,020
I think bad questions are things that
you could easily Google or find online,

2234
01:51:08,020 --> 01:51:14,220
and good questions are questions
that are not easily answered

2235
01:51:14,220 --> 01:51:16,530
online and actually
give you a picture of,

2236
01:51:16,530 --> 01:51:20,920
would you like to join this company,
and what is it like to work there?

2237
01:51:20,920 --> 01:51:26,310
So last advice I'll say is, don't worry.

2238
01:51:26,310 --> 01:51:30,510
Interviews are really,
really, really high variance.

2239
01:51:30,510 --> 01:51:35,250
There are great engineers who just
do super poorly on some interviews,

2240
01:51:35,250 --> 01:51:38,050
and that's just
unfortunately how it works.

2241
01:51:38,050 --> 01:51:42,060
So just because you didn't do
well on some number of interviews

2242
01:51:42,060 --> 01:51:43,830
doesn't mean at all
that you're not still

2243
01:51:43,830 --> 01:51:48,090
a great developer or data scientist or
designer or whatever path it might be.

2244
01:51:48,090 --> 01:51:50,910
Everyone has bad interviews sometimes.

2245
01:51:50,910 --> 01:51:52,080
Interviews are not fun.

2246
01:51:52,080 --> 01:51:54,640
I don't think I know anyone
who enjoys interviews, really

2247
01:51:54,640 --> 01:51:56,430
on either side of it.

2248
01:51:56,430 --> 01:52:00,690
But just view every interview you
do as an opportunity to get better.

2249
01:52:00,690 --> 01:52:04,170
If you don't pass an interview
or get rejected from a company,

2250
01:52:04,170 --> 01:52:06,208
just try to reflect on
that and think about what

2251
01:52:06,208 --> 01:52:07,500
went well, what didn't go well.

2252
01:52:07,500 --> 01:52:11,580
Maybe it was the type of problem,
maybe it was your communication.

2253
01:52:11,580 --> 01:52:14,910
Just take some time to learn
and grow from each experience,

2254
01:52:14,910 --> 01:52:17,640
but everyone is going to
go through a bad interview.

2255
01:52:17,640 --> 01:52:18,870
I've gone through plenty.

2256
01:52:18,870 --> 01:52:21,070
I'm sure everyone else on
this call has, as well.

2257
01:52:21,070 --> 01:52:24,030
And so don't worry if
that happens to you.

2258
01:52:24,030 --> 01:52:26,820
So last up, just some
takeaways from the talk--

2259
01:52:26,820 --> 01:52:28,620
one, have the basics down.

2260
01:52:28,620 --> 01:52:31,592
Going into any interview,
just know the basics

2261
01:52:31,592 --> 01:52:33,300
of the language you're
going to be using.

2262
01:52:33,300 --> 01:52:35,760
Be really comfortable in that setting.

2263
01:52:35,760 --> 01:52:36,810
Build your toolbox.

2264
01:52:36,810 --> 01:52:39,810
Be aware of the concepts we talked
about in those two mock interviews.

2265
01:52:39,810 --> 01:52:43,270
We looked at a string problem, we
looked at a binary search problem.

2266
01:52:43,270 --> 01:52:45,780
So those sort of concepts,
have them ready to go.

2267
01:52:45,780 --> 01:52:47,590
Do lots of practice problems.

2268
01:52:47,590 --> 01:52:49,560
The more you practice,
the more things you're

2269
01:52:49,560 --> 01:52:52,740
going to be able to pattern match
against, and the more comfortable

2270
01:52:52,740 --> 01:52:55,920
you're going to be in this really
unnatural interview setting.

2271
01:52:55,920 --> 01:52:58,860
Before you write any code,
just like Maria and Connor did,

2272
01:52:58,860 --> 01:53:01,350
communicate your
solution, explain the code

2273
01:53:01,350 --> 01:53:03,690
you're going to write,
and then write it out.

2274
01:53:03,690 --> 01:53:06,360
And then just like we saw in
those mock interviews, always

2275
01:53:06,360 --> 01:53:08,170
verbally run through your code.

2276
01:53:08,170 --> 01:53:10,320
Look for issues and
solve them proactively,

2277
01:53:10,320 --> 01:53:14,310
so the interviewer doesn't have
to tell you all of your bugs.

2278
01:53:14,310 --> 01:53:17,940
And I'll say one last time, all of
this is about the process, not about

2279
01:53:17,940 --> 01:53:18,690
the output.

2280
01:53:18,690 --> 01:53:22,980
Really focus on communicating well,
thinking critically, and getting

2281
01:53:22,980 --> 01:53:26,310
to a solution, rather than
writing absolutely perfect code

2282
01:53:26,310 --> 01:53:27,970
at the end of the day.

2283
01:53:27,970 --> 01:53:30,420
And so that's the end of the workshop.

2284
01:53:30,420 --> 01:53:32,820
Like I said, we'd love
to dive into questions

2285
01:53:32,820 --> 01:53:37,330
you have for me, as well as some
feedback on the mock interview.

2286
01:53:37,330 --> 01:53:41,367
And so I'll turn it over to David
to moderate questions and feedback.

2287
01:53:41,367 --> 01:53:43,950
DAVID MALAN: Yeah, Tommy, the
first question that came up was,

2288
01:53:43,950 --> 01:53:46,860
does Tommy have any feedback
on Connor's and Maria's

2289
01:53:46,860 --> 01:53:48,277
interviews, positive and negative?

2290
01:53:48,277 --> 01:53:49,693
TOMMY MACWILLIAMS: Yeah, for sure.

2291
01:53:49,693 --> 01:53:51,960
So I'd say both of those
interviews were positive.

2292
01:53:51,960 --> 01:53:55,530
I think, like I said, the most
important part was that communication.

2293
01:53:55,530 --> 01:53:57,450
And in both cases,
Connor and Maria started

2294
01:53:57,450 --> 01:54:01,860
by asking the sort of asking
clarifying questions in advance.

2295
01:54:01,860 --> 01:54:03,840
They made sure they
understood the problem.

2296
01:54:03,840 --> 01:54:07,360
In both cases, they made a list of
inputs and outputs, which I mentioned

2297
01:54:07,360 --> 01:54:09,540
is a really good way of framing things.

2298
01:54:09,540 --> 01:54:12,670
Also thinking out loud was
really, really important there.

2299
01:54:12,670 --> 01:54:16,260
I think in both cases, everybody
was constantly narrating.

2300
01:54:16,260 --> 01:54:18,840
They're never sitting in silence.

2301
01:54:18,840 --> 01:54:21,900
I think something else that I
like that Connor did specifically

2302
01:54:21,900 --> 01:54:24,060
was ask for a hint when he got stuck.

2303
01:54:24,060 --> 01:54:27,065
I think it would have been really
bad if-- imagine Connor was

2304
01:54:27,065 --> 01:54:29,940
at some mental wall and just didn't
see it and was sort of like stuck

2305
01:54:29,940 --> 01:54:30,750
and, like, I don't know.

2306
01:54:30,750 --> 01:54:32,340
Let me just sit there
and think about it.

2307
01:54:32,340 --> 01:54:35,548
Just asking for a hint, and then we were
able to get unstuck and and go right

2308
01:54:35,548 --> 01:54:37,090
down the right path.

2309
01:54:37,090 --> 01:54:41,130
So yeah, so I think those are the things
I would focus on for the interview.

2310
01:54:41,130 --> 01:54:43,110
I think in all cases, just making sure--

2311
01:54:43,110 --> 01:54:46,152
so we didn't really have enough time,
because it was intentionally short,

2312
01:54:46,152 --> 01:54:49,560
but just making sure the code is really
readable and clean and taking the time

2313
01:54:49,560 --> 01:54:51,775
to maybe create helper
functions or more readable

2314
01:54:51,775 --> 01:54:54,900
variable names-- we sort of omitted
that on purpose, just so the interviews

2315
01:54:54,900 --> 01:54:56,900
would be short, but that
would be the main thing

2316
01:54:56,900 --> 01:54:59,128
that I would do with more time.

2317
01:54:59,128 --> 01:55:00,920
DAVID MALAN: Another
question that came up,

2318
01:55:00,920 --> 01:55:03,620
especially for folks who
don't have the luxury of being

2319
01:55:03,620 --> 01:55:06,170
able to do mock interviews
with people like this,

2320
01:55:06,170 --> 01:55:09,650
are there recommended websites that
are good at preparing you for interview

2321
01:55:09,650 --> 01:55:11,270
questions for algorithms coding?

2322
01:55:11,270 --> 01:55:13,180
Is competitive programming helpful?

2323
01:55:13,180 --> 01:55:14,180
TOMMY MACWILLIAMS: Yeah.

2324
01:55:14,180 --> 01:55:16,790
I think there are lots
of great websites.

2325
01:55:16,790 --> 01:55:20,307
So one that I like a lot is called
LeetCode, so just leetcode.com.

2326
01:55:20,307 --> 01:55:21,890
They've got a bunch of problems there.

2327
01:55:21,890 --> 01:55:25,370
You'll probably find the ones that I
posted somewhere on LeetCode, as well,

2328
01:55:25,370 --> 01:55:26,665
because they have so many.

2329
01:55:26,665 --> 01:55:28,790
They're also great-- you
can write up your solution

2330
01:55:28,790 --> 01:55:32,570
and also see the solutions.

2331
01:55:32,570 --> 01:55:35,270
Another good one I like a
lot is called Project Euler.

2332
01:55:35,270 --> 01:55:37,760
Same thing, just gives
you practice problems.

2333
01:55:37,760 --> 01:55:42,450
You can go and implement them
in any language you want.

2334
01:55:42,450 --> 01:55:43,880
So I'd start there.

2335
01:55:43,880 --> 01:55:47,900
If you have a competitive programming
background, that's awesome.

2336
01:55:47,900 --> 01:55:50,090
I'm pretty bad at
algorithms problems myself.

2337
01:55:50,090 --> 01:55:52,795
So I think competitive
programming can be helpful.

2338
01:55:52,795 --> 01:55:54,920
I will say if you have done
competitive programming

2339
01:55:54,920 --> 01:55:57,200
before, the difference between
competitive programming

2340
01:55:57,200 --> 01:56:00,020
and interviewing is that
in competitive programming,

2341
01:56:00,020 --> 01:56:02,060
your variable names do not matter.

2342
01:56:02,060 --> 01:56:04,910
You're just trying to go super
fast, and oftentimes that

2343
01:56:04,910 --> 01:56:07,490
means being sloppy or
allocating these huge arrays.

2344
01:56:07,490 --> 01:56:09,953
You don't have to calculate
the links of things.

2345
01:56:09,953 --> 01:56:11,120
So definitely don't do that.

2346
01:56:11,120 --> 01:56:14,780
In a real interview, the readability
of your code matters a whole lot more.

2347
01:56:14,780 --> 01:56:17,720
So other sites, more
advanced sites, are things

2348
01:56:17,720 --> 01:56:23,275
like Top Coder, where there's other
sort of problems, practice problems.

2349
01:56:23,275 --> 01:56:26,150
They'll be a little more advanced,
but they're still a good practice.

2350
01:56:26,150 --> 01:56:27,900
DAVID MALAN: Another
question, Tommy, that

2351
01:56:27,900 --> 01:56:31,040
came from Pamika was, I wonder
if most interviewers would give

2352
01:56:31,040 --> 01:56:34,580
an interviewee a general problem
like this for entry level,

2353
01:56:34,580 --> 01:56:38,000
or if they would likely give a
more adaptive problem related

2354
01:56:38,000 --> 01:56:40,715
to the interviewee and
the job in question?

2355
01:56:40,715 --> 01:56:42,590
CONNOR LEGGETT: Yeah,
that's a good question,

2356
01:56:42,590 --> 01:56:45,260
and I think that's sort of what
we were getting at before about

2357
01:56:45,260 --> 01:56:46,970
the different types of interviews.

2358
01:56:46,970 --> 01:56:50,540
So I'd say, this is a really
common interview type.

2359
01:56:50,540 --> 01:56:53,962
What's nice is that it removes
a lot of the background,

2360
01:56:53,962 --> 01:56:55,920
because you can use
whatever language you want.

2361
01:56:55,920 --> 01:56:59,210
It doesn't require you to know some
particular framework or library.

2362
01:56:59,210 --> 01:57:01,400
You can solve this in
a whole bunch of ways.

2363
01:57:01,400 --> 01:57:03,570
That's the advantage
of a problem like this.

2364
01:57:03,570 --> 01:57:06,157
You may also see, in
addition to this, something

2365
01:57:06,157 --> 01:57:08,490
that's more specific to the
role that you're working on.

2366
01:57:08,490 --> 01:57:10,490
So if you're applying to
a web development role,

2367
01:57:10,490 --> 01:57:12,610
you might have a question
like, make a website.

2368
01:57:12,610 --> 01:57:14,360
So we didn't do a mock
interview for that,

2369
01:57:14,360 --> 01:57:18,050
but I'd say those are less common
than these really general questions.

2370
01:57:18,050 --> 01:57:22,250
But you could still run into them as
part of the whole interview process.

2371
01:57:22,250 --> 01:57:25,160
DAVID MALAN: And Tommy, for students
who aren't at Harvard or Yale

2372
01:57:25,160 --> 01:57:28,790
and might not be considering FAANG
companies like Facebook and Apple

2373
01:57:28,790 --> 01:57:31,790
and Amazon, Netflix, and
Google, is the kind of prep

2374
01:57:31,790 --> 01:57:36,980
that we're offering today focused on
those types of companies' interviews,

2375
01:57:36,980 --> 01:57:38,328
or tech more generally?

2376
01:57:38,328 --> 01:57:41,120
TOMMY MACWILLIAMS: Yeah, these are
really just focused on tech more

2377
01:57:41,120 --> 01:57:41,630
generally.

2378
01:57:41,630 --> 01:57:44,360
I think every company
will have its own process

2379
01:57:44,360 --> 01:57:45,740
and might do something different.

2380
01:57:45,740 --> 01:57:49,960
But the idea of being asked to
write some code in some setting

2381
01:57:49,960 --> 01:57:53,210
where someone is watching you and having
to think through your thought process

2382
01:57:53,210 --> 01:57:57,890
and think out loud, that's pretty
universal at a lot of companies.

2383
01:57:57,890 --> 01:58:00,890
And like I said, every company might
do something a little bit different

2384
01:58:00,890 --> 01:58:04,587
or have some more specific interviews,
maybe some online challenge

2385
01:58:04,587 --> 01:58:06,420
or something written
or something like that.

2386
01:58:06,420 --> 01:58:09,158
But holistically, you're thinking
about all the same things,

2387
01:58:09,158 --> 01:58:11,450
and all the prep work you do
is probably going to apply

2388
01:58:11,450 --> 01:58:13,995
to a really wide variety of interviews.

2389
01:58:13,995 --> 01:58:16,370
DAVID MALAN: And perhaps a
question for Maria or Connor--

2390
01:58:16,370 --> 01:58:19,445
how do you stop getting
nervous during your interviews?

2391
01:58:19,445 --> 01:58:21,320
CONNOR LEGGETT: So short
answer is you don't.

2392
01:58:21,320 --> 01:58:24,900
I get nervous before every interview.

2393
01:58:24,900 --> 01:58:26,930
But one thing that I--

2394
01:58:26,930 --> 01:58:32,480
like Tommy said before, practicing
really helps, and not just practicing

2395
01:58:32,480 --> 01:58:35,270
on your own or with
friends, but also practicing

2396
01:58:35,270 --> 01:58:37,550
in the form of other interviews.

2397
01:58:37,550 --> 01:58:43,260
So I've applied to a lot
of companies this year.

2398
01:58:43,260 --> 01:58:48,470
And so I have gotten to the point where
I'm doing a few interviews a week.

2399
01:58:48,470 --> 01:58:52,295
So the first week I was doing interviews
I was a lot more nervous than now,

2400
01:58:52,295 --> 01:58:54,760
where it seems like kind of a routine.

2401
01:58:54,760 --> 01:58:58,940
I know my introduction, I know
how to ask clarifying questions,

2402
01:58:58,940 --> 01:59:01,107
I know what questions I'm
going to ask interviewers

2403
01:59:01,107 --> 01:59:02,940
about their job, different
things like that.

2404
01:59:02,940 --> 01:59:05,090
So I would say that, just
like a lot of things,

2405
01:59:05,090 --> 01:59:07,280
the more you do it, the
less nervous you'll be.

2406
01:59:07,280 --> 01:59:09,290
MARIA ZLATKOVA: I would say repetition.

2407
01:59:09,290 --> 01:59:14,270
I also think, if you can, practice
with friends and kind of switch

2408
01:59:14,270 --> 01:59:16,730
up the interviewer-interviewee roles.

2409
01:59:16,730 --> 01:59:21,060
You can also involve some empathy
on the other side, as well.

2410
01:59:21,060 --> 01:59:24,020
But really just put yourself in
those situations and kind of simulate

2411
01:59:24,020 --> 01:59:26,330
the interview situation,
even when you're

2412
01:59:26,330 --> 01:59:28,070
solving practice problems on your own.

2413
01:59:28,070 --> 01:59:30,202
And like Tommy alluded
to earlier, don't just

2414
01:59:30,202 --> 01:59:32,660
be like, oh, I think I kind of
know how to do this and then

2415
01:59:32,660 --> 01:59:35,900
look at that solution and be like,
yeah, that's what I was going to do.

2416
01:59:35,900 --> 01:59:38,990
But actually time
yourself and go through it

2417
01:59:38,990 --> 01:59:42,517
and repeat as often as possible.

2418
01:59:42,517 --> 01:59:44,600
DAVID MALAN: And Tommy, a
technical question-- you

2419
01:59:44,600 --> 01:59:48,740
mentioned one or more times that
people should factor out common logic.

2420
01:59:48,740 --> 01:59:49,958
What did you mean by that?

2421
01:59:49,958 --> 01:59:53,000
TOMMY MACWILLIAMS: Something I say--
it's not like I say it all the time.

2422
01:59:53,000 --> 01:59:57,290
But oftentimes when you're writing
code, just generally, but also

2423
01:59:57,290 --> 02:00:02,540
in an interview, you might find that
some of the code you have is repeated.

2424
02:00:02,540 --> 02:00:06,110
Either it's repeated-- literally
the code is the same, like you're

2425
02:00:06,110 --> 02:00:09,560
doing the same thing to different
variables, or just parts of the logic

2426
02:00:09,560 --> 02:00:10,310
are repeated.

2427
02:00:10,310 --> 02:00:13,850
Maybe you're doing a
null check, or you're

2428
02:00:13,850 --> 02:00:17,370
doing some kind of string manipulation
that's similar in multiple parts.

2429
02:00:17,370 --> 02:00:19,430
So if you notice those
kind of similarities

2430
02:00:19,430 --> 02:00:24,020
or repeated logic in your code, try to
create a separate helper function that

2431
02:00:24,020 --> 02:00:28,010
just does that logic, and then call that
helper function from the main function

2432
02:00:28,010 --> 02:00:29,820
you've created for your interview.

2433
02:00:29,820 --> 02:00:33,198
So that process, I think it comes
from algebra, factoring things.

2434
02:00:33,198 --> 02:00:35,240
You could sort of factor
out the logic by saying,

2435
02:00:35,240 --> 02:00:36,907
here's these two things that are common.

2436
02:00:36,907 --> 02:00:39,840
Put it into some common function
that you call in both cases.

2437
02:00:39,840 --> 02:00:42,590
DAVID MALAN: And Tommy, a few folks
have asked whether they should

2438
02:00:42,590 --> 02:00:44,337
be commenting their code in interviews.

2439
02:00:44,337 --> 02:00:46,670
TOMMY MACWILLIAMS: Yeah,
that's an interesting question.

2440
02:00:46,670 --> 02:00:50,990
I will say your interviewer
preference might be different,

2441
02:00:50,990 --> 02:00:52,640
depending on the interviewer.

2442
02:00:52,640 --> 02:00:55,850
Me, personally, I wouldn't
spend a lot of time

2443
02:00:55,850 --> 02:00:58,550
on that, because you don't have
a lot of time in your interview.

2444
02:00:58,550 --> 02:01:01,902
You're already narrating really
naturally what your code is doing.

2445
02:01:01,902 --> 02:01:03,860
And so if you're just
typing out a comment that

2446
02:01:03,860 --> 02:01:05,610
matches exactly what
you're saying, that's

2447
02:01:05,610 --> 02:01:10,438
just describing what the code is doing,
it's probably not a great use of time.

2448
02:01:10,438 --> 02:01:12,480
Depending on your style
and your interview style,

2449
02:01:12,480 --> 02:01:15,300
maybe adding a quick docstring
to the function or just a comment

2450
02:01:15,300 --> 02:01:18,550
at the top of the function that explains
what it's going to do can be helpful.

2451
02:01:18,550 --> 02:01:22,620
But in general, I wouldn't spend a lot
of time commenting the code as you go,

2452
02:01:22,620 --> 02:01:24,060
given the setting.

2453
02:01:24,060 --> 02:01:25,590
DAVID MALAN: And we do have
time for a few more questions.

2454
02:01:25,590 --> 02:01:28,560
In fact, if you'd like to raise your
virtual hand in Zoom by clicking

2455
02:01:28,560 --> 02:01:31,020
the raised hand icon,
happy to call on some folks

2456
02:01:31,020 --> 02:01:33,210
if you have your video
and microphone on.

2457
02:01:33,210 --> 02:01:37,410
And Tommy, in the meantime, a couple
of folks asked, among them Aisha,

2458
02:01:37,410 --> 02:01:41,160
if you face a problem while writing your
code and can't think of the solution

2459
02:01:41,160 --> 02:01:43,890
even after a hint, what can you do?

2460
02:01:43,890 --> 02:01:46,200
TOMMY MACWILLIAMS: Yeah,
I will say, it happens.

2461
02:01:46,200 --> 02:01:48,798
I've hit many a wall in an interview.

2462
02:01:48,798 --> 02:01:51,840
I think, even if the interviewer gives
you a hint and you think about it,

2463
02:01:51,840 --> 02:01:55,050
I think it's super reasonable to
say, OK, I see what you're saying,

2464
02:01:55,050 --> 02:01:57,870
but I'm not sure how that
applies to the problem,

2465
02:01:57,870 --> 02:02:00,307
or I'm not sure I totally follow.

2466
02:02:00,307 --> 02:02:01,890
Can you give me some more information?

2467
02:02:01,890 --> 02:02:03,780
Basically, it's a conversation.

2468
02:02:03,780 --> 02:02:07,080
Your interviewer isn't
there to be a robot

2469
02:02:07,080 --> 02:02:09,850
or try to put you in a
situation that's tough.

2470
02:02:09,850 --> 02:02:11,290
They want you to succeed.

2471
02:02:11,290 --> 02:02:14,340
And so if they're trying to help
you, and they haven't helped you yet,

2472
02:02:14,340 --> 02:02:16,980
just keep asking for more
clarifying questions and more hints,

2473
02:02:16,980 --> 02:02:18,463
until eventually you get unstuck.

2474
02:02:18,463 --> 02:02:20,880
But sometimes I've been in
interviews where I'm like, hey,

2475
02:02:20,880 --> 02:02:22,530
I just didn't get it.

2476
02:02:22,530 --> 02:02:24,330
The interviewer had
to totally handhold me

2477
02:02:24,330 --> 02:02:26,470
through the problem, which
is usually what will happen.

2478
02:02:26,470 --> 02:02:28,530
They'll just give you
hints that are so obvious,

2479
02:02:28,530 --> 02:02:29,905
they're writing the code for you.

2480
02:02:29,905 --> 02:02:31,710
And then I got rejected, and that's--

2481
02:02:31,710 --> 02:02:32,430
it happens.

2482
02:02:32,430 --> 02:02:34,930
But the more you're making it
a conversation and a dialogue,

2483
02:02:34,930 --> 02:02:37,572
the more opportunities
you have to be helped.

2484
02:02:37,572 --> 02:02:39,780
DAVID MALAN: OK, and we have
a hand up from Yashasvi,

2485
02:02:39,780 --> 02:02:40,822
if I'm saying that right.

2486
02:02:40,822 --> 02:02:43,890
Yashasvi, if you'd like to say
hello and what your question is?

2487
02:02:43,890 --> 02:02:46,650
YASHASVI: I would like
to ask Tommy, sir,

2488
02:02:46,650 --> 02:02:51,840
what was the best thing one can do
to impress any interviewer, probably

2489
02:02:51,840 --> 02:02:54,470
in etiquette, or something like that?

2490
02:02:54,470 --> 02:02:55,470
TOMMY MACWILLIAMS: Sure.

2491
02:02:55,470 --> 02:02:59,320
So if I had one thing that would
impress an interviewer the most, for me,

2492
02:02:59,320 --> 02:03:02,470
the most important part of the
interview is communication.

2493
02:03:02,470 --> 02:03:05,262
So if I were to focus on one thing--
or the thing as an interviewer

2494
02:03:05,262 --> 02:03:07,095
that I come out of
interview, I'm like, wow,

2495
02:03:07,095 --> 02:03:10,180
that was great-- that thing is probably
because they communicated really,

2496
02:03:10,180 --> 02:03:10,860
really well.

2497
02:03:10,860 --> 02:03:13,110
Whether that was asking
clarifying questions,

2498
02:03:13,110 --> 02:03:15,342
just explaining their
approach or the way

2499
02:03:15,342 --> 02:03:18,300
they thought about the problem, that's
usually way more important to me

2500
02:03:18,300 --> 02:03:22,050
than, oh, cool, they use this crazy
Python built-in I didn't know existed.

2501
02:03:22,050 --> 02:03:25,080
Communication is way
more important than that.

2502
02:03:25,080 --> 02:03:28,320
DAVID MALAN: And a question from
Karthik, if you'd like to say hello,

2503
02:03:28,320 --> 02:03:30,980
where you're from, and
what your question is.

2504
02:03:30,980 --> 02:03:32,460
KARTHIK: I just have a question.

2505
02:03:32,460 --> 02:03:35,550
For a beginner level
or technical interview,

2506
02:03:35,550 --> 02:03:39,390
does clean code really matter?

2507
02:03:39,390 --> 02:03:46,740
The super most optimized code, really
it matters, or an optimum level

2508
02:03:46,740 --> 02:03:50,010
code or an ordinary level code?

2509
02:03:50,010 --> 02:03:51,397
I just want to know that.

2510
02:03:51,397 --> 02:03:53,730
TOMMY MACWILLIAMS: Yeah, so
I think the number one thing

2511
02:03:53,730 --> 02:03:55,570
that matters is the process.

2512
02:03:55,570 --> 02:03:57,680
So the way that you get
to that code and the way

2513
02:03:57,680 --> 02:03:59,430
you're reasoning and
communicating, that's

2514
02:03:59,430 --> 02:04:04,120
way more important than how
perfectly optimized your code was.

2515
02:04:04,120 --> 02:04:07,620
And then the other thing I'll say is
that the expectations the interviewer

2516
02:04:07,620 --> 02:04:10,560
has are going to vary, based
on your experience level.

2517
02:04:10,560 --> 02:04:15,570
So let's say I'm asking a question of
someone with 20 years of experience.

2518
02:04:15,570 --> 02:04:18,930
I'd expect their code to be
better than if I'm interviewing

2519
02:04:18,930 --> 02:04:20,430
someone for their first internship.

2520
02:04:20,430 --> 02:04:22,080
So I think it's going to vary.

2521
02:04:22,080 --> 02:04:25,117
As you're practicing, I'd say,
write the best code that you can.

2522
02:04:25,117 --> 02:04:27,450
Make sure that you're thinking
about, is the code clean?

2523
02:04:27,450 --> 02:04:30,720
But like I said, the process and
your thought process communication

2524
02:04:30,720 --> 02:04:33,780
is always going to be more
important than the code you write.

2525
02:04:33,780 --> 02:04:37,170
DAVID MALAN: Michael, would you like to
say hello and ask your question next?

2526
02:04:37,170 --> 02:04:41,070
MICHAEL: Yeah, hi, my name is
Michael, Fairfield, Connecticut.

2527
02:04:41,070 --> 02:04:43,690
So this is kind of--

2528
02:04:43,690 --> 02:04:45,690
I had to think a little
bit about this question,

2529
02:04:45,690 --> 02:04:48,810
because you guys covered so much, and
I felt like any technical questions

2530
02:04:48,810 --> 02:04:49,977
I had were already answered.

2531
02:04:49,977 --> 02:04:52,380
So anyway, I'll just ask you.

2532
02:04:52,380 --> 02:04:53,880
I'm a pretty big dude--

2533
02:04:53,880 --> 02:04:55,950
I'm six-four, 260 pounds.

2534
02:04:55,950 --> 02:05:00,210
Sometimes I don't have the nicest
face on, you know what I mean?

2535
02:05:00,210 --> 02:05:03,240
So what can I do when I get to
the interview and I'm nervous,

2536
02:05:03,240 --> 02:05:04,720
and I just look mean?

2537
02:05:04,720 --> 02:05:09,450
What are some things that maybe I
can do to make my interviewer feel

2538
02:05:09,450 --> 02:05:13,410
a little bit more comfortable
and know that we're here

2539
02:05:13,410 --> 02:05:15,032
trying to accomplish the same goals?

2540
02:05:15,032 --> 02:05:16,990
TOMMY MACWILLIAMS: Yeah,
it's a great question.

2541
02:05:16,990 --> 02:05:18,870
I think I'm also someone who--

2542
02:05:18,870 --> 02:05:22,290
I'm pretty bad at hiding my
emotions, which you could probably

2543
02:05:22,290 --> 02:05:25,860
see as I'm talking about resumes and
things I'm tired of seeing on them.

2544
02:05:25,860 --> 02:05:28,470
I think the biggest thing
is just, the interviewer

2545
02:05:28,470 --> 02:05:29,970
can tell when you're really nervous.

2546
02:05:29,970 --> 02:05:32,928
Sometimes I'll go on an interview,
and, I'm like, oh, man this person--

2547
02:05:32,928 --> 02:05:35,790
I can tell within 30 seconds
they're super nervous.

2548
02:05:35,790 --> 02:05:38,400
Me, as an interviewer,
I'll try to keep it light

2549
02:05:38,400 --> 02:05:40,600
and help calm you down
or make you less nervous.

2550
02:05:40,600 --> 02:05:43,657
Not every interviewer
can or wants to do that.

2551
02:05:43,657 --> 02:05:45,990
The number one thing is just
the more practice you have,

2552
02:05:45,990 --> 02:05:47,730
I think is really important.

2553
02:05:47,730 --> 02:05:51,142
For some people, having a glass
of water available to them

2554
02:05:51,142 --> 02:05:53,850
can be really helpful, so just
bringing a couple of water bottles

2555
02:05:53,850 --> 02:05:54,550
if it's on site.

2556
02:05:54,550 --> 02:05:56,770
I have a water bottle here,
just having that ready.

2557
02:05:56,770 --> 02:05:58,770
Taking a deep breath
before you start talking

2558
02:05:58,770 --> 02:06:03,000
or before you turn on that Zoom
call makes a huge, huge difference.

2559
02:06:03,000 --> 02:06:04,500
So I think little things like that--

2560
02:06:04,500 --> 02:06:07,318
everybody is a little
bit different about how

2561
02:06:07,318 --> 02:06:08,610
to make them feel less nervous.

2562
02:06:08,610 --> 02:06:11,820
But I think how nervous you
are on your first interview

2563
02:06:11,820 --> 02:06:14,970
and how nervous you are in, like,
your 50th is like such a big gap

2564
02:06:14,970 --> 02:06:18,210
that you just have to pay
that cost, go through a bunch.

2565
02:06:18,210 --> 02:06:20,970
And then for many people,
eventually, it starts

2566
02:06:20,970 --> 02:06:23,160
becoming less and less nerve-racking.

2567
02:06:23,160 --> 02:06:26,700
DAVID MALAN: Sarah, would you like to
say hello and ask your question next?

2568
02:06:26,700 --> 02:06:32,010
SARAH: Hello, and first of all, thank
you very much for your information.

2569
02:06:32,010 --> 02:06:37,380
And I want to ask, how
can we explain our job

2570
02:06:37,380 --> 02:06:43,200
gap before starting
this field, IT career,

2571
02:06:43,200 --> 02:06:46,980
I had a different experience
in different fields?

2572
02:06:46,980 --> 02:06:50,490
And then I gave a
break, like, five years.

2573
02:06:50,490 --> 02:06:57,000
After that five years, I'm starting
searching for a job in the IT field.

2574
02:06:57,000 --> 02:07:02,250
So how can I explain this
gap in my job experience?

2575
02:07:02,250 --> 02:07:02,752
Thank you.

2576
02:07:02,752 --> 02:07:04,710
TOMMY MACWILLIAMS: Yeah,
it's a great question.

2577
02:07:04,710 --> 02:07:07,740
And something I'll say is
that the experience you just

2578
02:07:07,740 --> 02:07:10,500
described of coming
from a different field

2579
02:07:10,500 --> 02:07:13,380
or taking a break in between
roles for some number of years,

2580
02:07:13,380 --> 02:07:15,810
it's becoming more and more common.

2581
02:07:15,810 --> 02:07:18,840
So maybe 10 years ago this was
something you really had to worry about.

2582
02:07:18,840 --> 02:07:22,270
People were only in engineering if
they had a degree in engineering.

2583
02:07:22,270 --> 02:07:25,770
I think now, so many people are
coming from other career paths

2584
02:07:25,770 --> 02:07:27,990
and going through boot
camps or online courses

2585
02:07:27,990 --> 02:07:31,710
or switching careers into
engineering that it almost

2586
02:07:31,710 --> 02:07:36,040
doesn't need to be explained as much
today as it would many years ago.

2587
02:07:36,040 --> 02:07:37,420
But I would just be honest.

2588
02:07:37,420 --> 02:07:38,910
I'd say, here's my background.

2589
02:07:38,910 --> 02:07:42,215
It was in some other field, could
be totally unrelated to software,

2590
02:07:42,215 --> 02:07:43,590
totally unrelated to engineering.

2591
02:07:43,590 --> 02:07:48,930
Talk about your experience there,
the impact you had, what you learned.

2592
02:07:48,930 --> 02:07:52,490
And then just talk about why you
want to switch to engineering

2593
02:07:52,490 --> 02:07:54,990
and what you've done so far and
why you're excited about it,

2594
02:07:54,990 --> 02:07:58,660
why you're excited to learn more
and help a company and have impact.

2595
02:07:58,660 --> 02:08:01,690
But I would say, my short answer is
I wouldn't worry too much about it.

2596
02:08:01,690 --> 02:08:04,890
And then to explain it, just be honest
about what you've done in the past

2597
02:08:04,890 --> 02:08:06,723
and why are you looking to switch.

2598
02:08:06,723 --> 02:08:08,640
DAVID MALAN: Misarg, if
I'm saying that right,

2599
02:08:08,640 --> 02:08:11,132
would you like to say hello
and ask your question.

2600
02:08:11,132 --> 02:08:13,590
MISARG: My name is Misarg
[INAUDIBLE],, and I'm from India.

2601
02:08:13,590 --> 02:08:17,200
And greetings to Tommy sir and
David sir and everyone else.

2602
02:08:17,200 --> 02:08:20,830
So my question, if you
get stuck on a problem,

2603
02:08:20,830 --> 02:08:22,830
what is the optimum
time to think over it

2604
02:08:22,830 --> 02:08:25,160
before stating that we
are unable to solve it?

2605
02:08:25,160 --> 02:08:27,160
TOMMY MACWILLIAMS: Yeah,
that's a good question.

2606
02:08:27,160 --> 02:08:32,190
So it's sort of like, at what
point do you ask for the hint?

2607
02:08:32,190 --> 02:08:36,190
I'll say one thing to keep mind is
the total duration of the interview,

2608
02:08:36,190 --> 02:08:38,850
so if you know you're in a
45-minute interview versus you're

2609
02:08:38,850 --> 02:08:40,530
in a two-hour interview.

2610
02:08:40,530 --> 02:08:43,320
My general heuristic is something
like three to five minutes.

2611
02:08:43,320 --> 02:08:46,740
If you've made no progress
in three to five minutes,

2612
02:08:46,740 --> 02:08:49,410
that's when I would
try to ask for a hint.

2613
02:08:49,410 --> 02:08:52,500
Sometimes just as you're
sort of thinking out loud,

2614
02:08:52,500 --> 02:08:54,540
you'll eventually run
out of things to say,

2615
02:08:54,540 --> 02:08:56,722
because you've run out
of things to think of.

2616
02:08:56,722 --> 02:08:59,430
And at that point, too, if you
find yourself being silent or just

2617
02:08:59,430 --> 02:09:02,830
not making any progress, think
for a minute or two and then say,

2618
02:09:02,830 --> 02:09:04,600
actually I'm not sure.

2619
02:09:04,600 --> 02:09:05,350
Can I have a hint?

2620
02:09:05,350 --> 02:09:08,183
So if you're thinking it's in the
ballpark of like 20 to 30 minutes,

2621
02:09:08,183 --> 02:09:10,320
let's say way lower
than that, but I also

2622
02:09:10,320 --> 02:09:12,860
would say much greater
than like 30 seconds,

2623
02:09:12,860 --> 02:09:14,610
because you'll give
yourself a few minutes

2624
02:09:14,610 --> 02:09:16,515
to actually process it and think.

2625
02:09:16,515 --> 02:09:18,640
DAVID MALAN: We have time
for a few more questions.

2626
02:09:18,640 --> 02:09:21,300
Do feel free to raise a virtual
hand and turn on your camera

2627
02:09:21,300 --> 02:09:22,680
if you'd like to ask on camera.

2628
02:09:22,680 --> 02:09:25,800
Or repaste your question in the
chat if we might have missed it.

2629
02:09:25,800 --> 02:09:28,590
We're doing our best to
field as many as we can.

2630
02:09:28,590 --> 02:09:32,470
Yuvraj, if I'm saying that right,
would you like to ask your question?

2631
02:09:32,470 --> 02:09:33,760
YUVRAG: So hi, guys.

2632
02:09:33,760 --> 02:09:37,590
My name is Raj, and I would
like to ask a question

2633
02:09:37,590 --> 02:09:40,560
from an entrepreneurial standpoint.

2634
02:09:40,560 --> 02:09:44,160
So a lot of companies
in the recent times

2635
02:09:44,160 --> 02:09:49,900
have changed the whole paradigm
of hiring people and all that.

2636
02:09:49,900 --> 02:09:54,780
So how do you expect that
it might change furthermore?

2637
02:09:54,780 --> 02:09:57,780
Even as a student if
I look at, things have

2638
02:09:57,780 --> 02:10:01,830
changed from going offline
doing interviews and all that,

2639
02:10:01,830 --> 02:10:05,400
to going online, due to
this pandemic and all that.

2640
02:10:05,400 --> 02:10:07,860
So do you think things
could change further,

2641
02:10:07,860 --> 02:10:11,827
or this is the top or
something like that?

2642
02:10:11,827 --> 02:10:14,910
TOMMY MACWILLIAMS: Yeah, it's a good
question, how interviews have evolved

2643
02:10:14,910 --> 02:10:16,590
and where they're going.

2644
02:10:16,590 --> 02:10:20,340
Certainly these virtual interviews
are way, way more common.

2645
02:10:20,340 --> 02:10:24,150
Many companies, even if
you could come on site,

2646
02:10:24,150 --> 02:10:25,890
just so much more
convenient to interview

2647
02:10:25,890 --> 02:10:29,670
someone remotely rather than
traveling six hours on a plane,

2648
02:10:29,670 --> 02:10:33,120
being tired at 9:00 AM the next day
for the interview, which is not fun.

2649
02:10:33,120 --> 02:10:34,412
So that's definitely a big one.

2650
02:10:34,412 --> 02:10:36,333
I think that's a trend that will stick.

2651
02:10:36,333 --> 02:10:38,250
I think many companies
will still do on-sites.

2652
02:10:38,250 --> 02:10:42,480
But I think before, I remember the
companies I've worked at and friends

2653
02:10:42,480 --> 02:10:46,170
have had a hard policy on, we have to
meet you in person before we hire you,

2654
02:10:46,170 --> 02:10:49,470
just cannot join the team
unless we meet you in person.

2655
02:10:49,470 --> 02:10:51,570
I don't know why, but that was a policy.

2656
02:10:51,570 --> 02:10:54,470
I don't think you'll see very
many things like that anymore.

2657
02:10:54,470 --> 02:10:56,220
I think another thing
that's been changing

2658
02:10:56,220 --> 02:11:02,500
is that many companies are asking
more of these practical questions,

2659
02:11:02,500 --> 02:11:04,920
especially to folks
with more experience.

2660
02:11:04,920 --> 02:11:06,270
And I think that's good and bad.

2661
02:11:06,270 --> 02:11:10,050
I think for many folks coming
out of university or new grads,

2662
02:11:10,050 --> 02:11:12,120
asking them a practical
question in many cases

2663
02:11:12,120 --> 02:11:15,030
is worse, just because they haven't
ever written production code

2664
02:11:15,030 --> 02:11:16,620
or launched a production website.

2665
02:11:16,620 --> 02:11:18,600
And so asking somebody
to do that, it actually

2666
02:11:18,600 --> 02:11:20,017
doesn't give you that much signal.

2667
02:11:20,017 --> 02:11:23,605
And asking an algorithms question to a
student who's taken algorithms classes,

2668
02:11:23,605 --> 02:11:26,230
and they're really good at that,
can actually be higher signal.

2669
02:11:26,230 --> 02:11:28,560
And so where I see this
shift coming is that,

2670
02:11:28,560 --> 02:11:31,140
if you have the experience
of building these systems,

2671
02:11:31,140 --> 02:11:32,400
we're going to ask
you about it and maybe

2672
02:11:32,400 --> 02:11:33,858
ask you to build something similar.

2673
02:11:33,858 --> 02:11:35,340
And if you don't, then we're not.

2674
02:11:35,340 --> 02:11:37,983
Where I think a few years ago,
there's this blanket everybody

2675
02:11:37,983 --> 02:11:40,650
has to do whiteboard questions,
because whiteboard questions are

2676
02:11:40,650 --> 02:11:41,580
the be-all end-all.

2677
02:11:41,580 --> 02:11:43,497
And I'd say they have a purpose.

2678
02:11:43,497 --> 02:11:46,080
A lot of people are like,
whiteboard questions are always bad,

2679
02:11:46,080 --> 02:11:46,747
they're useless.

2680
02:11:46,747 --> 02:11:48,180
I don't subscribe to that at all.

2681
02:11:48,180 --> 02:11:51,360
I've interviewed so many students who
do better on these than they do on

2682
02:11:51,360 --> 02:11:52,980
build a website questions.

2683
02:11:52,980 --> 02:11:54,870
So they have their place,
and I think they're

2684
02:11:54,870 --> 02:11:59,005
becoming more of a specialized thing
than a blanket part of the process.

2685
02:11:59,005 --> 02:12:00,880
DAVID MALAN: And Tommy,
a couple of questions

2686
02:12:00,880 --> 02:12:02,755
that have come up a few
times now in the chat

2687
02:12:02,755 --> 02:12:06,580
relate to folks who either don't have
much work experience, because they're

2688
02:12:06,580 --> 02:12:09,190
just coming out of college,
or they're older and have

2689
02:12:09,190 --> 02:12:11,950
gaps in their resume for some reason.

2690
02:12:11,950 --> 02:12:14,440
What kinds of personal
technical projects

2691
02:12:14,440 --> 02:12:16,340
should people in those
situations take on?

2692
02:12:16,340 --> 02:12:18,490
And when it comes to
languages, given that they

2693
02:12:18,490 --> 02:12:21,340
might have a choice of languages,
which languages should they

2694
02:12:21,340 --> 02:12:22,550
be picking up for interviews?

2695
02:12:22,550 --> 02:12:24,550
TOMMY MACWILLIAMS: Yeah,
that's a good question.

2696
02:12:24,550 --> 02:12:26,200
I'll start with the second one.

2697
02:12:26,200 --> 02:12:29,300
I would pick up languages
that are more popular.

2698
02:12:29,300 --> 02:12:34,210
So I think that takes the form of Python
or JavaScript, would probably be my top

2699
02:12:34,210 --> 02:12:35,200
two.

2700
02:12:35,200 --> 02:12:38,590
Those are languages you could use
to build a little web application

2701
02:12:38,590 --> 02:12:41,680
or visualization or a
command line application.

2702
02:12:41,680 --> 02:12:46,510
I wouldn't try to pick up as a first
language something like Haskell or Rust

2703
02:12:46,510 --> 02:12:49,652
or some of these more
sort of less mainstream

2704
02:12:49,652 --> 02:12:51,110
and have a higher barrier to entry.

2705
02:12:51,110 --> 02:12:53,740
So Python or JavaScript are
good ones to start with.

2706
02:12:53,740 --> 02:12:56,500
And then for a personal
project, I would say build

2707
02:12:56,500 --> 02:12:58,090
something that's interesting to you.

2708
02:12:58,090 --> 02:13:02,358
My best advice to people is that
everyone in their day-to-day

2709
02:13:02,358 --> 02:13:04,150
has some problem that
they're dealing with.

2710
02:13:04,150 --> 02:13:11,230
Maybe it's taking notes in class or
organizing some files in their laptop.

2711
02:13:11,230 --> 02:13:14,470
Everyone sort of has these little
problems that they deal with.

2712
02:13:14,470 --> 02:13:17,620
Just write a little app to
solve them for yourself.

2713
02:13:17,620 --> 02:13:20,050
I remember one of the
first things I built

2714
02:13:20,050 --> 02:13:22,300
was this little weather
visualizer, because I

2715
02:13:22,300 --> 02:13:24,250
didn't like the weather.com things.

2716
02:13:24,250 --> 02:13:27,280
I'm like, oh, let me use their API
and just build a little weather app.

2717
02:13:27,280 --> 02:13:30,910
And I just had that in my room
where I could wake up and check

2718
02:13:30,910 --> 02:13:32,862
my little weather app.

2719
02:13:32,862 --> 02:13:35,320
There's thousands of weather
apps out there, but who cares?

2720
02:13:35,320 --> 02:13:37,240
It was like a problem that I
thought was kind of interesting,

2721
02:13:37,240 --> 02:13:38,300
and I went and solved it.

2722
02:13:38,300 --> 02:13:41,540
So everybody has those little
things that are interesting to them.

2723
02:13:41,540 --> 02:13:43,600
So pick something that's
interesting to you

2724
02:13:43,600 --> 02:13:46,600
and write a little command line
application or web application

2725
02:13:46,600 --> 02:13:48,310
to solve it.

2726
02:13:48,310 --> 02:13:51,280
DAVID MALAN: A few
final questions here--

2727
02:13:51,280 --> 02:13:53,530
Sarah, would you like
to ask your question

2728
02:13:53,530 --> 02:13:55,870
and tell us where you are in the world?

2729
02:13:55,870 --> 02:14:00,610
SARAH: Hi, I'm Sarah from Algeria.

2730
02:14:00,610 --> 02:14:04,520
And I was wondering, do you think
I could ask all the questions that

2731
02:14:04,520 --> 02:14:07,420
are in my mind about the work and
the company to my interviewer,

2732
02:14:07,420 --> 02:14:11,350
even after I know that I failed
with my code and writing the code

2733
02:14:11,350 --> 02:14:12,850
and solving the problem?

2734
02:14:12,850 --> 02:14:15,512
Do I have the right to ask them or not?

2735
02:14:15,512 --> 02:14:18,220
TOMMY MACWILLIAMS: Oh, at the end
of the interview when they ask,

2736
02:14:18,220 --> 02:14:20,290
do you have any questions
for me, even if you

2737
02:14:20,290 --> 02:14:22,000
know that the interview didn't go well?

2738
02:14:22,000 --> 02:14:22,930
SARAH: Yes.

2739
02:14:22,930 --> 02:14:24,430
TOMMY MACWILLIAMS: Yeah, I think so.

2740
02:14:24,430 --> 02:14:26,800
I think you should
always-- they're still

2741
02:14:26,800 --> 02:14:29,830
interested in giving you a good
experience with the company.

2742
02:14:29,830 --> 02:14:32,590
Even though you didn't pass
this interview, who knows?

2743
02:14:32,590 --> 02:14:36,970
You might come back a year or two later,
interview again, and then do fantastic.

2744
02:14:36,970 --> 02:14:40,600
And so if the interviewer can give
you a good picture of the company now,

2745
02:14:40,600 --> 02:14:42,985
that might help you say,
oh, I will go reapply again,

2746
02:14:42,985 --> 02:14:44,860
because I really like
the person I talked to,

2747
02:14:44,860 --> 02:14:46,943
and I've learned a lot
since that first interview.

2748
02:14:46,943 --> 02:14:50,502
So yeah, I definitely say when you
have a bad interview experience,

2749
02:14:50,502 --> 02:14:51,460
it happens to everyone.

2750
02:14:51,460 --> 02:14:53,620
Try not to get too down on yourself.

2751
02:14:53,620 --> 02:14:54,730
When they ask, do you
have any question for me,

2752
02:14:54,730 --> 02:14:56,563
I'd still ask the
prepared questions you had

2753
02:14:56,563 --> 02:15:00,175
about their day-to-day or the things
that they change about the company.

2754
02:15:00,175 --> 02:15:01,550
So I wouldn't change any of that.

2755
02:15:01,550 --> 02:15:03,010
And then who knows?

2756
02:15:03,010 --> 02:15:06,345
You could see that same interviewer
again a few years later.

2757
02:15:06,345 --> 02:15:08,470
DAVID MALAN: And Tomas,
would you like to say hello

2758
02:15:08,470 --> 02:15:11,330
and where you're from,
along with your question?

2759
02:15:11,330 --> 02:15:13,090
TOMAS: What's up, guys?

2760
02:15:13,090 --> 02:15:15,430
I've enjoyed a lot,
and I'd like to know,

2761
02:15:15,430 --> 02:15:18,130
sometimes when it comes
to an interview, there

2762
02:15:18,130 --> 02:15:20,860
are companies that ask
you about your salary.

2763
02:15:20,860 --> 02:15:23,920
By the way, I'm from
Venezuela at the [INAUDIBLE]..

2764
02:15:23,920 --> 02:15:27,460
So when it comes to
interviews, there are companies

2765
02:15:27,460 --> 02:15:28,870
that ask you about the salary.

2766
02:15:28,870 --> 02:15:33,430
So is there a way to know
what would be the best?

2767
02:15:33,430 --> 02:15:39,700
Because I think this was
not so covering in this.

2768
02:15:39,700 --> 02:15:45,280
TOMMY MACWILLIAMS: Yeah, so how to think
about salary and negotiation generally?

2769
02:15:45,280 --> 02:15:48,880
TOMAS: A way to know what
would be the best amount or--

2770
02:15:48,880 --> 02:15:52,300
because sometimes we could
ask too much or too little.

2771
02:15:52,300 --> 02:15:57,320
And are we even supposed to answer that,
or we can just avoid that question?

2772
02:15:57,320 --> 02:15:59,320
TOMMY MACWILLIAMS: Yeah,
this one's a tough one,

2773
02:15:59,320 --> 02:16:03,850
because it's going to vary a lot
by role and country and company.

2774
02:16:03,850 --> 02:16:08,260
So I can try to give some general
advice that may or may not

2775
02:16:08,260 --> 02:16:10,780
be a little too US-specific,
just because I'm not

2776
02:16:10,780 --> 02:16:13,660
super familiar with all the
different international conventions.

2777
02:16:13,660 --> 02:16:16,420
But something that is
certainly applicable

2778
02:16:16,420 --> 02:16:20,410
is before you go into a role, try
to get a general sense for what

2779
02:16:20,410 --> 02:16:23,170
the market is paying for that role.

2780
02:16:23,170 --> 02:16:26,898
There's lots of sites online you can
look at, whether they be Glassdoor--

2781
02:16:26,898 --> 02:16:28,690
there's another one
that's in the US called

2782
02:16:28,690 --> 02:16:34,120
Levels FYI, where people just
publish, here are the ranges of what

2783
02:16:34,120 --> 02:16:35,799
I'm making at this role.

2784
02:16:35,799 --> 02:16:37,886
The company might also
publish it in advance.

2785
02:16:37,886 --> 02:16:40,719
There are some sites like Angel
Lists that start-up used to recruit,

2786
02:16:40,719 --> 02:16:43,090
and it's sort of a required
field, is how much are you

2787
02:16:43,090 --> 02:16:46,870
going to be offering for this role?

2788
02:16:46,870 --> 02:16:52,030
So yeah, just do a little
bit of research in advance.

2789
02:16:52,030 --> 02:16:54,580
I'd say try to give a range too.

2790
02:16:54,580 --> 02:16:58,642
I think some people are like, I must
make at least n-amount of dollars

2791
02:16:58,642 --> 02:16:59,559
or I'm not interested.

2792
02:16:59,559 --> 02:17:01,476
I think that's usually
not the right approach,

2793
02:17:01,476 --> 02:17:03,942
but giving sort of a
general range, also sort

2794
02:17:03,942 --> 02:17:06,400
of talking about what you're
currently making in your role,

2795
02:17:06,400 --> 02:17:09,940
and what you'd ideally like to
be making more than that, having

2796
02:17:09,940 --> 02:17:14,388
your current compensation can
be kind of a baseline there.

2797
02:17:14,388 --> 02:17:16,180
So I think, yeah, my
general advice is just

2798
02:17:16,180 --> 02:17:19,525
to try to be honest and try to
do some research in advance.

2799
02:17:19,525 --> 02:17:21,400
And then that can help
you avoid this outcome

2800
02:17:21,400 --> 02:17:25,158
where you're asking for something
that's way too low or way too high.

2801
02:17:25,158 --> 02:17:27,700
But yeah, every company is going
to be a little bit different

2802
02:17:27,700 --> 02:17:28,780
in how they handle this.

2803
02:17:28,780 --> 02:17:31,330
Some companies want you to disclose
what you're looking for first.

2804
02:17:31,330 --> 02:17:34,360
Some companies say, here's the role and
the level we're going to level you at.

2805
02:17:34,360 --> 02:17:36,180
Here's what people in
that role and level make.

2806
02:17:36,180 --> 02:17:37,847
It's going to be a little bit different.

2807
02:17:37,847 --> 02:17:40,062
But the research part you can always do.

2808
02:17:40,062 --> 02:17:42,520
DAVID MALAN: I think we have
time for a few more questions.

2809
02:17:42,520 --> 02:17:44,650
Hope folks will forgive if
we don't get to everyone.

2810
02:17:44,650 --> 02:17:47,275
We also want to leave a moment
at the end for Maria and Connor,

2811
02:17:47,275 --> 02:17:50,360
if you're amenable, to offer just
some final thoughts of your own.

2812
02:17:50,360 --> 02:17:53,469
But first, Aisha, if I'm saying that
right, would you like to say hello,

2813
02:17:53,469 --> 02:17:56,110
where you're from, and your question?

2814
02:17:56,110 --> 02:17:59,559
AISHA: I am from
Pakistan, and it's Aisha.

2815
02:17:59,559 --> 02:18:01,670
DAVID MALAN: Aisha.

2816
02:18:01,670 --> 02:18:02,170
AISHA: Yeah.

2817
02:18:02,170 --> 02:18:05,389
And first of all, I have to say
that this lesson has been great.

2818
02:18:05,389 --> 02:18:09,980
I did not know most of the things,
and some things really surprised me.

2819
02:18:09,980 --> 02:18:12,760
So I absolutely enjoyed this lesson.

2820
02:18:12,760 --> 02:18:19,000
And my question is does it matter,
the style of the interview?

2821
02:18:19,000 --> 02:18:22,030
Does that differ if you have a
different level of experience

2822
02:18:22,030 --> 02:18:26,049
or your level of your
education or something

2823
02:18:26,049 --> 02:18:30,730
like that or the level of what you are,
like in CS, if I'm an absolute beginner

2824
02:18:30,730 --> 02:18:33,969
or I have done really less
experience, one year or two years,

2825
02:18:33,969 --> 02:18:36,129
does it matter the style
of the interviewer,

2826
02:18:36,129 --> 02:18:39,129
the interviewer itself,
or the assignment?

2827
02:18:39,129 --> 02:18:41,180
TOMMY MACWILLIAMS: Yeah,
it's a great question,

2828
02:18:41,180 --> 02:18:43,400
so how things vary based on experience.

2829
02:18:43,400 --> 02:18:47,170
So the general setting
tends to be the same.

2830
02:18:47,170 --> 02:18:50,889
Even people with a lot of experience
will still ask a question of the form,

2831
02:18:50,889 --> 02:18:54,129
write a function that takes this
as input and takes that as output.

2832
02:18:54,129 --> 02:18:58,487
I think the expectations will vary a
bit for someone who's really new to CS.

2833
02:18:58,487 --> 02:19:01,570
The problem we ask might be a little
easier, a little more straightforward

2834
02:19:01,570 --> 02:19:02,049
to solve.

2835
02:19:02,049 --> 02:19:04,007
For somebody who has more
experience, might ask

2836
02:19:04,007 --> 02:19:05,667
something a little bit harder.

2837
02:19:05,667 --> 02:19:07,750
And then I mentioned before,
something that you'll

2838
02:19:07,750 --> 02:19:10,660
see as you have more experience
is more questions that

2839
02:19:10,660 --> 02:19:13,780
are a little more practical, telling
me about how you've architected

2840
02:19:13,780 --> 02:19:18,190
some large-scale system in the past or
to build something from scratch that

2841
02:19:18,190 --> 02:19:22,360
would be production-ready, where for
students, they're usually more asking

2842
02:19:22,360 --> 02:19:26,170
questions about the things that
you've covered in your courses.

2843
02:19:26,170 --> 02:19:28,990
Many people take a data structures
or an algorithms course,

2844
02:19:28,990 --> 02:19:31,690
and so we'll ask a question about
data structures or algorithms.

2845
02:19:31,690 --> 02:19:34,990
I think in general, the ideal
interview and what a lot of interviews

2846
02:19:34,990 --> 02:19:37,990
do-- and hopefully they're all
doing-- is looking at your experience,

2847
02:19:37,990 --> 02:19:41,870
seeing what you've done in the past,
and then asking questions based on that.

2848
02:19:41,870 --> 02:19:45,200
So if you've never built a
large-scale backend system,

2849
02:19:45,200 --> 02:19:47,200
asking you to build a
large-scale backend system

2850
02:19:47,200 --> 02:19:48,283
doesn't really make sense.

2851
02:19:48,283 --> 02:19:51,460
I'd rather ask questions that we're
going to get a lot of signal out of

2852
02:19:51,460 --> 02:19:54,760
or I'm going to understand, are you a
good fit for the role I'm hiring for.

2853
02:19:54,760 --> 02:19:57,052
And if I just ask you something
where you have no idea,

2854
02:19:57,052 --> 02:19:59,663
then I'm actually not
getting any information.

2855
02:19:59,663 --> 02:20:01,830
DAVID MALAN: Andrew, we saw
your hand go up earlier.

2856
02:20:01,830 --> 02:20:04,775
Would you like to say hello
and ask your question?

2857
02:20:04,775 --> 02:20:15,320
ANDREW: My question is, can I get the
job after completing university only?

2858
02:20:15,320 --> 02:20:18,320
TOMMY MACWILLIAMS: Tommy, can one get
a job after completing university?

2859
02:20:18,320 --> 02:20:20,350
TOMMY MACWILLIAMS: Yeah, definitely.

2860
02:20:20,350 --> 02:20:23,470
There's lots of job
opportunities that are especially

2861
02:20:23,470 --> 02:20:26,170
designed for folks who are
just coming out of university

2862
02:20:26,170 --> 02:20:28,360
and haven't had any jobs before.

2863
02:20:28,360 --> 02:20:31,510
Internships can also be a great way
of getting a role if you haven't

2864
02:20:31,510 --> 02:20:34,330
had any experience yet, so
definitely lots of great roles

2865
02:20:34,330 --> 02:20:36,875
you can get just out of college.

2866
02:20:36,875 --> 02:20:40,000
DAVID MALAN: Nice, and can we go next
to Shalom, if you'd like to say hello

2867
02:20:40,000 --> 02:20:42,120
and where you're from
with your question?

2868
02:20:42,120 --> 02:20:48,430
SHALOM: My name is Shalom, and I am
from South Africa, currently in Beijing.

2869
02:20:48,430 --> 02:20:53,510
My question is, I studied chemical
engineering and technology.

2870
02:20:53,510 --> 02:20:56,680
So this is really new to me.

2871
02:20:56,680 --> 02:21:00,850
I recently started my course on CS50.

2872
02:21:00,850 --> 02:21:05,978
My question is, what do I study
after completing that course?

2873
02:21:05,978 --> 02:21:08,020
TOMMY MACWILLIAMS: Yeah,
that's a great question,

2874
02:21:08,020 --> 02:21:09,550
after CS50, where do you go next?

2875
02:21:09,550 --> 02:21:12,770
I'm sure David has
thoughts on this, as well.

2876
02:21:12,770 --> 02:21:15,430
But I would say in general,
for interviewing specifically,

2877
02:21:15,430 --> 02:21:20,050
any course that has data structures
or algorithms written in it,

2878
02:21:20,050 --> 02:21:22,240
those can be really, really
helpful, so the courses

2879
02:21:22,240 --> 02:21:26,650
where you're going to
learn what a binary tree is

2880
02:21:26,650 --> 02:21:29,590
and how to implement a hash table,
some of which we cover in CS50,

2881
02:21:29,590 --> 02:21:32,480
but there's more advanced data
structures and algorithms.

2882
02:21:32,480 --> 02:21:36,097
So I'd say those are the most useful
courses you can take for interviewing.

2883
02:21:36,097 --> 02:21:38,680
And then maybe David wants to
answer the more general question

2884
02:21:38,680 --> 02:21:40,185
of where to go next.

2885
02:21:40,185 --> 02:21:42,310
DAVID MALAN: Yeah, we often
recommend that students

2886
02:21:42,310 --> 02:21:47,000
explore another type of
programming, functional programming,

2887
02:21:47,000 --> 02:21:49,450
object-oriented programming,
keywords like those,

2888
02:21:49,450 --> 02:21:52,420
as well as a good course on
data structures and algorithms,

2889
02:21:52,420 --> 02:21:53,530
for instance.

2890
02:21:53,530 --> 02:21:55,900
I'm biased, too, because I
work with a whole team here

2891
02:21:55,900 --> 02:21:58,900
that produces courses that can be
taken both before and after CS50.

2892
02:21:58,900 --> 02:22:02,200
But Maria, can we perhaps ask
you, having been in the real world

2893
02:22:02,200 --> 02:22:06,010
as an engineer for some time, what you
found most helpful after an intro class

2894
02:22:06,010 --> 02:22:08,320
like CS50?

2895
02:22:08,320 --> 02:22:14,350
MARIA ZLATKOVA: Yeah, so I think really
trying to get a breadth of classes

2896
02:22:14,350 --> 02:22:17,080
out there and exploring different
interests would be good.

2897
02:22:17,080 --> 02:22:20,560
When I first took CS50,
some of my next classes

2898
02:22:20,560 --> 02:22:25,900
that I took were systems, operating
programming, data structures

2899
02:22:25,900 --> 02:22:28,930
and algorithms, data visualization.

2900
02:22:28,930 --> 02:22:33,160
And I found a lot of the more
visual and design-oriented classes

2901
02:22:33,160 --> 02:22:38,540
super interesting, so definitely got
a chance to focus in on that, as well.

2902
02:22:38,540 --> 02:22:43,180
But I definitely would say that
exploring a breadth of things

2903
02:22:43,180 --> 02:22:47,110
so that you can really decide what
you want to hone in on and what really

2904
02:22:47,110 --> 02:22:49,297
interests you is always a good idea.

2905
02:22:49,297 --> 02:22:51,880
DAVID MALAN: And I think Tommy's
emphasis on personal projects

2906
02:22:51,880 --> 02:22:54,790
earlier is important, because
we see among online students

2907
02:22:54,790 --> 02:22:58,720
a real tendency to take one course, then
another course, then another course,

2908
02:22:58,720 --> 02:23:01,000
and it's never really
clear when it ends.

2909
02:23:01,000 --> 02:23:03,178
And it doesn't need to
end, but at some point,

2910
02:23:03,178 --> 02:23:05,470
you should really leave the
nest, so to speak, and take

2911
02:23:05,470 --> 02:23:07,960
on some personal project,
pursue some work project,

2912
02:23:07,960 --> 02:23:11,050
so that you feel like you're
actually now applying the skills

2913
02:23:11,050 --> 02:23:13,330
that you've learned in classrooms.

2914
02:23:13,330 --> 02:23:15,708
Some final questions-- can
we go next to Remington,

2915
02:23:15,708 --> 02:23:18,250
if you'd like to ask your question
and say where you're from?

2916
02:23:18,250 --> 02:23:19,840
REMINGTON: My name is Remington.

2917
02:23:19,840 --> 02:23:23,842
I'm currently in the UK in quarantine.

2918
02:23:23,842 --> 02:23:25,300
It's lovely to join you guys today.

2919
02:23:25,300 --> 02:23:26,890
It was super helpful.

2920
02:23:26,890 --> 02:23:29,020
I don't mean to focus
too much on the resumes,

2921
02:23:29,020 --> 02:23:32,390
but it's another resume question.

2922
02:23:32,390 --> 02:23:36,550
So if I'm a student,
non-CS major, and I've

2923
02:23:36,550 --> 02:23:40,720
had some internships that
are not really tech-related,

2924
02:23:40,720 --> 02:23:43,870
is it more valuable to add those
to kind of fill out my resume,

2925
02:23:43,870 --> 02:23:46,930
or should I really try to
do more personal projects

2926
02:23:46,930 --> 02:23:49,040
and then add those
into my resume instead?

2927
02:23:49,040 --> 02:23:51,250
So which one would be more
valuable, is my question.

2928
02:23:51,250 --> 02:23:54,250
TOMMY MACWILLIAMS: Yeah, I would say
if I had to pick between those two,

2929
02:23:54,250 --> 02:23:55,720
I would say the latter.

2930
02:23:55,720 --> 02:23:59,830
So having personal projects
that are related to the roles

2931
02:23:59,830 --> 02:24:01,663
that you are applying
to are really helpful.

2932
02:24:01,663 --> 02:24:04,247
So for instance, if you're
applying to a web development role,

2933
02:24:04,247 --> 02:24:07,240
showing some websites you've
built is really helpful.

2934
02:24:07,240 --> 02:24:10,092
And I'd say even if you only
have a few, that's totally fine.

2935
02:24:10,092 --> 02:24:12,550
Fill out the rest of your space
with your prior experience.

2936
02:24:12,550 --> 02:24:15,790
But an interviewer will
definitely love to see something

2937
02:24:15,790 --> 02:24:18,280
that's related to the role
that you're applying to more

2938
02:24:18,280 --> 02:24:20,572
than the past experience you do have.

2939
02:24:20,572 --> 02:24:23,530
So I'd say projects first, then fill
out the rest with prior experience

2940
02:24:23,530 --> 02:24:24,030
you do have.

2941
02:24:24,030 --> 02:24:25,730
And definitely include that, as well.

2942
02:24:25,730 --> 02:24:29,080
DAVID MALAN: And Shaurya, if I'm saying
that right, would you like to say hello

2943
02:24:29,080 --> 02:24:30,073
and your question?

2944
02:24:30,073 --> 02:24:31,240
SHAURYA: My name is Shaurya.

2945
02:24:31,240 --> 02:24:32,530
I'm from India.

2946
02:24:32,530 --> 02:24:35,890
My question was, is keeping
a smile on your face

2947
02:24:35,890 --> 02:24:38,530
while the interview is
not going that much well,

2948
02:24:38,530 --> 02:24:43,720
so does that affect any
impact on the interviewer?

2949
02:24:43,720 --> 02:24:45,910
TOMMY MACWILLIAMS:
Yeah, I'd say whether--

2950
02:24:45,910 --> 02:24:48,850
everyone knows that interviews
can be really stressful,

2951
02:24:48,850 --> 02:24:54,580
and so I don't think I've ever expected
anyone to be like, totally happy

2952
02:24:54,580 --> 02:24:57,320
and not stress the whole time.

2953
02:24:57,320 --> 02:24:59,230
So I would say the best--

2954
02:24:59,230 --> 02:25:00,980
it all comes down to communication.

2955
02:25:00,980 --> 02:25:03,040
I think as long as you're
communicating clearly,

2956
02:25:03,040 --> 02:25:05,410
and you're having a good
dialogue with your interviewer,

2957
02:25:05,410 --> 02:25:09,010
and you're able to share
your ideas and listen

2958
02:25:09,010 --> 02:25:11,873
to feedback from your
interviewer and hear their hints,

2959
02:25:11,873 --> 02:25:13,540
I think that's the most important thing.

2960
02:25:13,540 --> 02:25:17,260
But everybody's really
nervous, and I wouldn't

2961
02:25:17,260 --> 02:25:20,170
recommend anyone trying to
fake how they're feeling

2962
02:25:20,170 --> 02:25:24,490
or pretend to be feeling a certain
way, when the reality is that everybody

2963
02:25:24,490 --> 02:25:25,900
knows that it's nerve-racking.

2964
02:25:25,900 --> 02:25:28,720
And so interviewers don't expect
you to be doing that anyway,

2965
02:25:28,720 --> 02:25:30,580
and it all just comes
down to communication.

2966
02:25:30,580 --> 02:25:32,330
As long as the
communication is effective,

2967
02:25:32,330 --> 02:25:33,790
that's the most important thing.

2968
02:25:33,790 --> 02:25:38,020
DAVID MALAN: Let's go lastly to Pawan
Kumar, if I'm saying that right.

2969
02:25:38,020 --> 02:25:38,635
My apologies.

2970
02:25:38,635 --> 02:25:41,690
PAWAN KUMAR: So thank you so
much for this opportunity.

2971
02:25:41,690 --> 02:25:44,710
So my question is more on
the system design level.

2972
02:25:44,710 --> 02:25:49,840
So yeah, somehow we did so much
hard work for some brilliant years,

2973
02:25:49,840 --> 02:25:53,920
and somehow we were able to
break the coding interviews.

2974
02:25:53,920 --> 02:25:59,020
But then we have another piece
called system design, where someone

2975
02:25:59,020 --> 02:25:59,870
with my experience--

2976
02:25:59,870 --> 02:26:04,370
I have 10 years experience,
I live in the United States.

2977
02:26:04,370 --> 02:26:07,040
Though I have experience when
it comes to their questions

2978
02:26:07,040 --> 02:26:11,920
on how to design a Facebook, how
do you design a Twitter, yeah,

2979
02:26:11,920 --> 02:26:14,680
we design things in our
site, but it doesn't

2980
02:26:14,680 --> 02:26:16,480
mean it's going to be that level.

2981
02:26:16,480 --> 02:26:20,260
So can you please give me some idea of
how do I prepare for the system design

2982
02:26:20,260 --> 02:26:21,010
interviews?

2983
02:26:21,010 --> 02:26:24,135
TOMMY MACWILLIAMS: Yeah, for sure, I'll
give some quick thoughts on systems

2984
02:26:24,135 --> 02:26:25,900
design interviews.

2985
02:26:25,900 --> 02:26:28,150
So generally speaking,
what we're looking

2986
02:26:28,150 --> 02:26:32,380
for in a systems design interview
is often something architectural.

2987
02:26:32,380 --> 02:26:35,890
So for example, you mentioned
how you build the Facebook news

2988
02:26:35,890 --> 02:26:36,710
feed, for instance.

2989
02:26:36,710 --> 02:26:39,760
So we're not looking
for precisely like, I

2990
02:26:39,760 --> 02:26:42,880
would use 0 mq with a
fan-out architecture

2991
02:26:42,880 --> 02:26:45,010
and a multilayered
[INAUDIBLE] tier for caching.

2992
02:26:45,010 --> 02:26:47,260
If you could say that,
that's great, but we're

2993
02:26:47,260 --> 02:26:50,380
more concerned about
architectural things, like here

2994
02:26:50,380 --> 02:26:52,510
is what the schema of
the database would be,

2995
02:26:52,510 --> 02:26:54,530
here's how rows are
written to the database,

2996
02:26:54,530 --> 02:26:57,640
here's where the caching layers
are, what's being cached,

2997
02:26:57,640 --> 02:26:59,140
here's what's being read from those.

2998
02:26:59,140 --> 02:27:01,640
So it's almost like if the goal
of a system design interview

2999
02:27:01,640 --> 02:27:05,230
is to have a diagram
out on a whiteboard,

3000
02:27:05,230 --> 02:27:08,110
either verbally or
literally, that explains

3001
02:27:08,110 --> 02:27:12,220
all of the major components of
the system and how they interact.

3002
02:27:12,220 --> 02:27:15,460
And that's much more
important than picking

3003
02:27:15,460 --> 02:27:18,580
the individual library or the
sort of implementation details

3004
02:27:18,580 --> 02:27:20,140
within that diagram.

3005
02:27:20,140 --> 02:27:23,770
I think if you can speak to
technologies you have used in the past

3006
02:27:23,770 --> 02:27:26,438
and say, in the past we use
Kafka for message passing,

3007
02:27:26,438 --> 02:27:28,480
and here's some things I
liked about it, but here

3008
02:27:28,480 --> 02:27:30,522
are some ways we're kind
of broke down and things

3009
02:27:30,522 --> 02:27:33,370
like that, that experience
is really valuable.

3010
02:27:33,370 --> 02:27:35,830
But for a systems
design interview, we're

3011
02:27:35,830 --> 02:27:38,900
almost never looking for
some really precise answer.

3012
02:27:38,900 --> 02:27:41,410
The other thing we're also
not necessarily looking for

3013
02:27:41,410 --> 02:27:44,560
is exactly what part of
the system you design.

3014
02:27:44,560 --> 02:27:47,770
In many cases, it's really
intentionally ambiguous.

3015
02:27:47,770 --> 02:27:51,660
One that I've got in the past is,
how would you make Google Maps?

3016
02:27:51,660 --> 02:27:53,560
End of question, turn it over to me.

3017
02:27:53,560 --> 02:27:54,480
And that could mean anything, right?

3018
02:27:54,480 --> 02:27:55,890
That could be the front
end, that could be,

3019
02:27:55,890 --> 02:27:58,050
like, how do you implement
all the scrolling.

3020
02:27:58,050 --> 02:28:01,500
It could be, how do you
store all of the map?

3021
02:28:01,500 --> 02:28:05,923
There's tiles of different
elevation levels and resolutions.

3022
02:28:05,923 --> 02:28:06,840
How do you store that?

3023
02:28:06,840 --> 02:28:08,463
It could be, how do you do--

3024
02:28:08,463 --> 02:28:09,630
Google Maps includes routes.

3025
02:28:09,630 --> 02:28:11,190
How do you do route
planning and traffic?

3026
02:28:11,190 --> 02:28:12,930
And it was sort of
intentionally ambiguous,

3027
02:28:12,930 --> 02:28:13,860
and the interviewer was
like, oh, I don't know,

3028
02:28:13,860 --> 02:28:15,660
whatever you want to design, do it.

3029
02:28:15,660 --> 02:28:19,000
You've got an hour, let's
just see what we can do.

3030
02:28:19,000 --> 02:28:22,165
And so I keep that in mind, as well,
that sometimes it might feel like

3031
02:28:22,165 --> 02:28:25,207
the interviewer doesn't have anything
in mind, and it's because we don't.

3032
02:28:25,207 --> 02:28:27,168
We just want to see
where you gravitate to,

3033
02:28:27,168 --> 02:28:29,460
because you're naturally
going to gravitate, hopefully,

3034
02:28:29,460 --> 02:28:31,627
towards the thing where you
have the most experience

3035
02:28:31,627 --> 02:28:33,543
and where you can design the best thing.

3036
02:28:33,543 --> 02:28:35,460
So yeah, so really focus
on that architecture.

3037
02:28:35,460 --> 02:28:37,680
Focus on what are the
components of the system,

3038
02:28:37,680 --> 02:28:40,350
how do those components interact,
what are the abstraction

3039
02:28:40,350 --> 02:28:41,970
barriers between those systems.

3040
02:28:41,970 --> 02:28:45,540
Then if you can talk about
implementation details like frameworks

3041
02:28:45,540 --> 02:28:50,160
or libraries or patterns you've used
in the past, then you can do so.

3042
02:28:50,160 --> 02:28:53,987
But we're really mostly interested
in architectural decisions.

3043
02:28:53,987 --> 02:28:55,820
DAVID MALAN: And Tommy,
[INAUDIBLE] question

3044
02:28:55,820 --> 02:29:00,137
was about what should you do when you
get frustrated and stuck with coding?

3045
02:29:00,137 --> 02:29:02,720
TOMMY MACWILLIAMS: Yeah, I think
in an interview, like I said,

3046
02:29:02,720 --> 02:29:04,910
it happens to everybody.

3047
02:29:04,910 --> 02:29:07,220
It'll happen to you,
and it's happened to me.

3048
02:29:07,220 --> 02:29:10,220
I think just taking a
deep breath and taking

3049
02:29:10,220 --> 02:29:14,390
just a step back, just 10 seconds, take
a step back and think and calm yourself

3050
02:29:14,390 --> 02:29:15,770
down as best as you can.

3051
02:29:15,770 --> 02:29:17,920
Take a glass of water if they have one.

3052
02:29:17,920 --> 02:29:20,780
If they don't, bring
one so you have one.

3053
02:29:20,780 --> 02:29:22,700
And then just ask for a hint.

3054
02:29:22,700 --> 02:29:26,720
I think a lot of interviewees,
when they get frustrated,

3055
02:29:26,720 --> 02:29:29,150
they sort of get angry
or sort of get mad.

3056
02:29:29,150 --> 02:29:31,100
And I think it certainly might happen.

3057
02:29:31,100 --> 02:29:34,850
Just try to, as best you
can, try to take a step back

3058
02:29:34,850 --> 02:29:40,090
and ask for help in a way that's sort
of keeping the conversation light

3059
02:29:40,090 --> 02:29:41,890
and not becoming overly angry.

3060
02:29:41,890 --> 02:29:45,482
But it certainly can happen and
has happened to me and people

3061
02:29:45,482 --> 02:29:46,940
I've been interviewing in the past.

3062
02:29:46,940 --> 02:29:50,680
So yeah, I would say if you get
frustrated, just ask for a hint,

3063
02:29:50,680 --> 02:29:53,890
take a deep breath, and remind
yourself that everybody fails

3064
02:29:53,890 --> 02:29:55,587
these things, that it's not just you.

3065
02:29:55,587 --> 02:29:57,920
DAVID MALAN: I have folks who
still have their hands up.

3066
02:29:57,920 --> 02:30:01,760
Well, forgive if we didn't get to all
questions, but in our remaining minutes

3067
02:30:01,760 --> 02:30:04,400
here, just wanted to
give Maria and Connor

3068
02:30:04,400 --> 02:30:07,790
and then Tommy the final
word with some final advice.

3069
02:30:07,790 --> 02:30:08,600
Maria?

3070
02:30:08,600 --> 02:30:12,590
MARIA ZLATKOVA: Yeah, I really
would just say, don't be worried

3071
02:30:12,590 --> 02:30:14,180
or scared when you mess up.

3072
02:30:14,180 --> 02:30:17,900
I have messed up so many interviews, and
I sometimes still think back to them.

3073
02:30:17,900 --> 02:30:21,050
And I kind of find them
funny now thinking back,

3074
02:30:21,050 --> 02:30:25,010
because I really just put so much
pressure on myself that it even

3075
02:30:25,010 --> 02:30:28,410
prevented me from doing well
and showing what I knew well

3076
02:30:28,410 --> 02:30:30,860
and actually thinking
clearly in the moment.

3077
02:30:30,860 --> 02:30:33,590
And we said this a lot,
but the best way to do

3078
02:30:33,590 --> 02:30:35,690
that is to really just
practice and put yourself

3079
02:30:35,690 --> 02:30:39,540
in those high-stress situations
to get yourself more used to them.

3080
02:30:39,540 --> 02:30:43,760
And we believe in all of you,
especially in coming to this,

3081
02:30:43,760 --> 02:30:46,770
I think you're taking the
best steps that you can.

3082
02:30:46,770 --> 02:30:49,730
So it really is just about
taking this and these guidelines

3083
02:30:49,730 --> 02:30:53,760
that we shared today and just practicing
and really mastering everything.

3084
02:30:53,760 --> 02:30:57,500
So thanks so much for spending
the past few hours with us.

3085
02:30:57,500 --> 02:30:58,750
DAVID MALAN: Thank you, Maria.

3086
02:30:58,750 --> 02:31:00,865
And Connor?

3087
02:31:00,865 --> 02:31:02,990
CONNOR LEGGETT: Yeah, I
definitely agree with that.

3088
02:31:02,990 --> 02:31:05,440
I would say the biggest thing is to--

3089
02:31:05,440 --> 02:31:09,850
even though it's very difficult, to
try and be as relaxed as you can.

3090
02:31:09,850 --> 02:31:11,750
Try to have a good outlook on it.

3091
02:31:11,750 --> 02:31:17,110
One thing that I think I've found
out over the last couple of years

3092
02:31:17,110 --> 02:31:19,300
that I had wished I
had known going into it

3093
02:31:19,300 --> 02:31:23,740
is that a lot of the interview
process is a bit random.

3094
02:31:23,740 --> 02:31:28,480
So you might get a question that
you feel really confident about

3095
02:31:28,480 --> 02:31:31,420
or a question that you just
have no idea what to do.

3096
02:31:31,420 --> 02:31:34,790
And there can also be interviews where
you feel like you did really well,

3097
02:31:34,790 --> 02:31:37,690
and then you don't get into
the next round, or interviews

3098
02:31:37,690 --> 02:31:41,390
that you feel like you bombed, and
then you do get into the next round.

3099
02:31:41,390 --> 02:31:44,290
So I think that's kind
of helpful to know,

3100
02:31:44,290 --> 02:31:47,990
because even if you struggle
with the question for a while,

3101
02:31:47,990 --> 02:31:50,410
it's not a good idea to
give up on that interview.

3102
02:31:50,410 --> 02:31:54,360
You want to keep trying, keep
interacting with your interviewer,

3103
02:31:54,360 --> 02:31:57,087
and you never know
what's going to happen.

3104
02:31:57,087 --> 02:31:58,670
DAVID MALAN: Well, thank you, as well.

3105
02:31:58,670 --> 02:32:00,730
And thank you to all of the
team here behind the scenes

3106
02:32:00,730 --> 02:32:02,938
who've been making this
possible and certainly Tommy,

3107
02:32:02,938 --> 02:32:07,000
who prepared so much of today's
materials, along with Maria and Connor.

3108
02:32:07,000 --> 02:32:10,330
Tommy, final advice you'd
like to leave folks with?

3109
02:32:10,330 --> 02:32:13,150
TOMMY MACWILLIAMS: Final
advice, I'll say don't give up.

3110
02:32:13,150 --> 02:32:15,130
I'll never forget my
first interview where I

3111
02:32:15,130 --> 02:32:16,960
was interviewing at a big tech company.

3112
02:32:16,960 --> 02:32:18,850
And we're in a building
with an elevator,

3113
02:32:18,850 --> 02:32:21,940
and the two people from the interview,
my interviewer and someone else

3114
02:32:21,940 --> 02:32:23,440
in the company were in the elevator.

3115
02:32:23,440 --> 02:32:26,080
And they were talking about
how bad I did in the interview,

3116
02:32:26,080 --> 02:32:27,430
like, this kid was such a joke.

3117
02:32:27,430 --> 02:32:29,805
And I walked into the elevator
and then took the elevator

3118
02:32:29,805 --> 02:32:32,650
down 10 floors with them as
they just like stared at me,

3119
02:32:32,650 --> 02:32:34,807
and everybody knew what was happening.

3120
02:32:34,807 --> 02:32:35,890
So I'll never forget that.

3121
02:32:35,890 --> 02:32:37,300
It happens to everybody.

3122
02:32:37,300 --> 02:32:38,750
Definitely don't give up.

3123
02:32:38,750 --> 02:32:42,220
You can fail a ton of interviews
before you eventually succeed,

3124
02:32:42,220 --> 02:32:43,240
and that's totally OK.

3125
02:32:43,240 --> 02:32:44,900
Everybody goes through that.

3126
02:32:44,900 --> 02:32:46,900
So yeah, you might feel
like this is impossible.

3127
02:32:46,900 --> 02:32:48,220
Who can possibly do this?

3128
02:32:48,220 --> 02:32:48,910
You can.

3129
02:32:48,910 --> 02:32:50,867
Just don't give up,
and you'll get there.

3130
02:32:50,867 --> 02:32:52,450
DAVID MALAN: Thank you so much, Tommy.

3131
02:32:52,450 --> 02:32:53,750
Thank you for joining us today.

3132
02:32:53,750 --> 02:32:56,970
We'll see you online
for CS50 again soon.

3133
02:32:56,970 --> 02:32:58,222