Hackernoon logo'How I got a $500K job at Facebook (as a software engineer)' - TechLead by@techlead

'How I got a $500K job at Facebook (as a software engineer)' - TechLead

Author profile picture

@techleadPatrick Shyu

YouTuber (500k subs), Ex-Google/ex-Facebook Tech Lead.

Ex-Google TechLead explains the interview process at Facebook for staff software engineer. (FYI, this video is a little more technical in nature.)

Check out my new project CoderPro (https://www.coderpro.com) for 50+ coding video explanations to popular interview questions for software engineers.

Join my Discord: https://discord.gg/pFUBUtE

๐ŸŽ‰ Party up:

Disclosure: Some links are affiliate links to products. I may receive a small commission for purchases made through these links. #techlead

Video Transcript

(Note: This transcript is auto-generated by YouTube and may not be 100% accurate.)

00:00

hey welcome back to coffee time with

00:01

your host X Google X Facebook tech lead

00:03

and it occurred to me that I never

00:05

really explained how I got my job at

00:07

Facebook as a staff software engineer

00:09

making five hundred thousand dollars per

00:11

year

00:11

today I thought I would tell you about

00:13

how that interview process went I landed

00:15

this offer about one and a half years

00:17

ago when I was working over at Google as

00:19

a tech lead I might remind you before

00:22

all the drama about my wife leaving me

00:24

happened before I became unemployable

00:26

for bashing Facebook but this may be a

00:28

path that you would like to pursue as

00:30

well at some point in your future maybe

00:32

not now maybe not ever really I wanted

00:35

to explain why might even give you an to

00:37

survive if you know the information if

00:39

you know knowledge if you can pass these

00:40

interviews you should be qualified for

00:42

the job it would otherwise be like

00:44

saying students I say Harvard somehow

00:46

have some secret bit of information that

00:49

is non-public and then they only tell

00:50

that information to their friends and

00:52

then only those people are able to get

00:53

into Harvard and I feel that this

00:55

information disparity is not fair it's

00:57

not fair for people who don't have those

00:59

professional network connections who may

01:01

be from underprivileged communities so

01:03

that's my perspective about giving you

01:05

guys as much information as you can to

01:07

pass these coding interviews

01:08

all right so Facebook generally has

01:10

three different types of interviews you

01:12

can usually tell which one you may be at

01:13

they have coding systems design and

01:16

behavior so the coding focuses on just

01:19

your average sale each code the type of

01:21

questions and what you have to solve

01:22

some coding exercise and that's really

01:24

what's a I'll go pro which I talked

01:26

about it's going to help you get through

01:27

and a lot of junior engineers they

01:29

stumble on this portion or they just

01:31

miss out on a lot of these problems the

01:33

systems design interview is not so

01:35

focused on coding and in fact they'll

01:36

probably get a little bit mad if you

01:38

start coding too much they want you to

01:39

talk about how these pieces of code

01:41

interact with each other and that

01:43

overall architecture of it and then

01:45

there's the behavior portion in which

01:47

they tend to value your ability to

01:48

handle conflict build consensus lead

01:51

projects push impact and overall have

01:53

good collaboration and communication

01:54

skills because overall programming is a

01:57

team sport for myself since I was

01:59

interviewing for a senior role I had one

02:01

coding question two systems design

02:03

interviews and then one behavioral

02:05

portion and in each of these they

02:06

generally do continue to ask you some

02:08

coding questions so it's not like they

02:10

just ask you one let me tell you about

02:11

the coding questions that I had though

02:13

and there's a good thing I took notes on

02:15

these so I can tell you the exact

02:16

questions I was asked and you can think

02:18

about how you may solve some of these as

02:20

well you can just pause the video as we

02:22

go along here the first question is

02:24

given to duplicate binary trees so you

02:27

got tree a and three B and you're given

02:30

a node and tree a find that same

02:32

corresponding node in those B so you can

02:35

think about how you may solve this by

02:37

remember my approach was I would try to

02:39

traverse one tree find the note and then

02:41

traverse the second tree and see if I

02:43

can find that same note but then I

02:45

realized that there was no good way to

02:47

compare the two nodes the interviewer

02:49

kind of stopped me at that point I had

02:51

to step back and think about it and the

02:53

ultimate approach which I think is the

02:55

right one is to just reverse both at the

02:56

same time and when you reach the node in

02:59

a then you know you've reached that same

03:00

node in the second binary tree as well

03:03

but the main trick here is that even

03:05

though most people may have been taught

03:06

to traverse one binary tree at a time

03:08

doing both treat reversals at the same

03:11

time sort of test your mastery of this

03:13

area now the funny thing is for a

03:15

standard interview I think this would be

03:16

the only question but I actually ended

03:18

up solving like three questions in the

03:20

single session so the second question

03:22

they asked was given an array like one

03:25

zero zero two five zero move all of the

03:27

zeros to the right to the end of the

03:29

array and again this isn't necessarily

03:32

so difficult but the challenge is to do

03:34

it in like a single pass linear time

03:36

algorithm make it very efficient and

03:38

come up with the proper time space

03:40

analysis and one brute force technique

03:42

is you can go through the array find the

03:45

first zero and move it to the end go

03:46

through the array find the next zero

03:47

move it to the end and you just keep

03:49

moving these values and that's probably

03:51

going to take quadratic time n square

03:53

time because you got to go through n

03:55

iterations and for each iteration you

03:57

may have to shift and more elements the

03:59

more optimal solution can be done in

04:01

linear time and constant space you keep

04:03

in the index to the last item in the

04:05

array and anytime you find the zero you

04:07

just swap that element with the last

04:09

item and then you move that last index

04:11

one left and you just keep going like

04:13

that and then you can process through

04:15

the whole range it's a pretty ridiculous

04:17

question but you just got to play that

04:18

game

04:19

the next question is check of a string

04:21

of parentheses valid a very common

04:23

question you always get asked this type

04:25

of thing you can either solve that using

04:27

like a

04:27

Hunter but if there are multiple

04:29

different types of parentheses then you

04:30

can use like a stack and when you

04:32

encounter and open the in parentheses

04:34

you push that on to the stack we

04:35

encounter a closing parenthesis you pop

04:37

that off of the stack and you check that

04:39

they match and then there's a follow-up

04:41

question which was to remove all

04:42

unbalanced parentheses another question

04:45

was you're given two binary strings and

04:47

you need to add them manually

04:48

essentially just do a duration using a

04:50

carryover variable that can be done in

04:52

linear time constant space and we go

04:54

over a question similar to this over an

04:56

algo Pro except that they're linked

04:58

lists instead so it's a little bit

04:59

trickier that way and then the last

05:01

question is you are given an array of

05:03

objects small medium and large and that

05:05

you have to sort them in place and you

05:08

know this was a mix of using the proper

05:10

data structures because they weren't

05:11

numbers there's like small medium and

05:13

large objects in there and then coming

05:15

up with the algorithm to sort this

05:17

efficiently the optimal algorithm would

05:19

use no space and it would be linear time

05:21

just do the entire sort in one iteration

05:24

one past I remember the original

05:26

approach I did was first I had a bucket

05:28

of small items and then medium and large

05:30

items combined and I would first go

05:32

through in one pass and move all the

05:34

small items to the beginning and then

05:36

we'll go through a second pass I move

05:37

all the large items to the end and just

05:39

sort the medium and large items so this

05:41

is actually a two pass algorithm so that

05:43

covers the coding and all software

05:45

engineers jr. through senior needs a

05:47

known coding you're gonna be smoked on

05:48

this no matter what level you're at

05:50

which is why it's important to keep your

05:52

skill sharp with algo procom now the two

05:55

systems design questions I got were both

05:57

pretty similar and it's totally going to

05:59

depend on the area that you're in right

06:01

for me I'm an iOS mobile developer so it

06:04

was about building an iOS app

06:05

specifically it was number one design a

06:08

flight booking app and number two design

06:10

a simple version of Instagram so both of

06:12

these questions are about having

06:14

multiple pages or surfaces in an app

06:16

there could be dozens of teams working

06:18

on these and then architecting the app

06:20

structure in the way that multiple

06:21

people can be working on this and then

06:23

handling things like caching offline

06:25

usage event navigation scroll

06:27

performance layout removing inter

06:29

dependencies between different pages for

06:31

the flight booking application the idea

06:33

is that you first search for a set of

06:34

flights you see a list you choose a

06:36

flight and then you can buy the flight

06:37

similarly for the Instagram one is like

06:40

you

06:41

have a bunch of photos then you can

06:42

click on the photo you see the details

06:44

and it's like a two-page app you can

06:47

really go in any direction that you want

06:48

in these systems design interviews is

06:50

super open for me the areas I drove into

06:53

talking more about we're handling

06:54

offline usage scenarios how do you catch

06:57

that stuff

06:58

separating dependencies between surfaces

07:00

so for example the flight booking app

07:02

you may have a payments team that is

07:04

handling the payment view for buying the

07:06

airplane ticket but then you have

07:08

another team that's handling search and

07:10

then another team that handles

07:11

displaying the list of flights and it's

07:14

like how do you structure the

07:15

application in a way such that all of

07:17

these can be independent of one another

07:19

and yet they're able to function

07:20

together and here I will talk about

07:22

having different event handlers

07:24

separating modules out such that they're

07:26

not dependent upon one another and

07:28

coming up with that module dependency

07:29

graph and I had to sort of back step

07:31

they're taking their feedback and during

07:33

the interview I actually realized an

07:35

even better architectural solution that

07:37

we hadn't figured out over at Google

07:38

that Facebook had been implementing and

07:40

they probably just appreciated that I

07:43

was able to collaborate with them to

07:44

solve that right there and then there

07:46

are also questions about holistically

07:48

structuring the app in a way that could

07:50

be suitable for both offline and online

07:52

use and then making the right

07:53

abstractions that say the network layer

07:55

and then I also talked about the UI as

07:57

well so that covers what a mobile

07:59

systems design interview may look like

08:01

for a web it will probably be very

08:02

different for example they'll probably

08:04

be about load balancers databases

08:06

master-slave replication sharding CD ins

08:09

and all of that I think what really

08:11

helped me though was really being more

08:12

interested in other parts of the code

08:14

base rather than whatever I have been

08:16

assigned right like usually people are

08:18

saying to work on one small feature and

08:19

they never take a step back and look at

08:21

how their feature interacts with

08:23

everything else and get an overview of

08:25

the entire code base that they're

08:26

working on if you can take a step back

08:29

and take a look at that that could

08:30

prepare you well for any systems design

08:32

interview that may be coming your way at

08:34

some point maybe now or in the future

08:36

and then lots of other the behavior

08:39

questions which we cover in far more

08:40

depth over in our other program Tech

08:43

interview program where we covered the

08:44

entire interview process start to finish

08:47

but the behavioral questions they really

08:49

dig and they don't accept your first

08:52

answer in fact they're trained not to so

08:54

if they ask you like what

08:55

the most challenging part about this and

08:56

you give some shallow answer like oh

08:58

yeah working with this person and then

09:00

it overcame this challenge they know

09:02

that that first answer

09:03

it's probably like a BS answer that you

09:05

have prepared so they really dig deep

09:06

and this is where it helps to really

09:09

have good examples

09:10

I prepared a number of stories myself

09:11

about the only with conflict tough

09:13

situations at work and I think that for

09:15

this area it also helps you just have

09:16

some good concrete examples which is

09:19

another reason that in whatever role

09:21

you're in it still helps to be pushing

09:23

yourself challenging yourself to do more

09:25

such that it can be something you can

09:26

talk about as you're interviewing for

09:28

your future roles so there you have it I

09:30

remember leaving the interview feeling

09:31

quite positive about it and then two

09:33

weeks later I would receive the offer

09:34

five hundred thousand dollar salary and

09:36

there was still kind of a hard decision

09:38

for me to leave Google to take up that

09:40

offer because I was quite comfortable

09:42

over at Google in fact I felt I was

09:44

probably getting a little bit too

09:45

comfortable just sitting back the idea

09:47

of learning a whole new codebase new

09:49

technologies it just seemed really

09:51

exciting to me at the time so I made

09:52

that switch turned out later to be one

09:55

of the biggest mistakes of my life but

09:57

that's another story

09:58

so that good for me it was a little bit

10:01

more of a technical video but I gotta

10:03

keep my street cred up as the tech lead

10:04

if you liked the video give it a like

10:06

and subscribe and I'll see you next time

10:08

Thanks bye

Tags

Join Hacker Noon

Create your free account to unlock your custom reading experience.