CS career megathread / AMA

[suralin]

i was a phil major at an unranked liberal arts college, discovered TLS, switched to comp sci late, and have been a software / data engineer at FB since may '17 (3-4 yrs of experience total)

i know an embarrassing amount about ~big tech~ / CS / coding careers, especially about bootstrapping from a shitty starting point, and am v down to help anybody who is curious or wants to career PIVOT

also AMA

e: other tech / CS ppl are v welcome to contribute as well

[riot]

I'm glad this was the first post in the legal employment forum

[suralin]

I'm glad this was the first post in the legal employment forum

#movefast

[wizzy]

I'm glad this was the first post in the legal employment forum

Me too. Totally on team "everybody should try out coding in high school/college to see if they like it/have a knack for it."

Also cool that we have a sealocust equivalent for CS. Hopefully we can get some additional STEM posters and finance people as well.

[MJB]

For lawyers not looking to make a career change, is coding worth trying out just for tech literacy purposes? A bunch of people in chambers were thinking about that after reading one of those adoring pieces on William "The Judge Who Codes" Alsup.

[Kali]

Can also contribute. Analytic consultant focused on, well, analytics and data science at a systems integrator. Also a law school grad but transitioned away after no-offered from biglaw summer program.

[wizzy]

Can also contribute. Analytic consultant focused on, well, analytics and data science at a systems integrator. Also a law school grad but transitioned away after no-offered from biglaw summer program.

How long did it take you to learn R, Python, SQL, SAS, etc.? Which did you learn first?

[Kali]

Can also contribute. Analytic consultant focused on, well, analytics and data science at a systems integrator. Also a law school grad but transitioned away after no-offered from biglaw summer program.

How long did it take you to learn R, Python, SQL, SAS, etc.? Which did you learn first?

SQL is stupid easy. It took me maybe two months to be good enough to query out my info *reliably*. I focused on R because I am not a software engineer and will never put shit into production. Most of my data analysis is ad hoc, so R made sense. If I had to do it all over again, I might have done Python simply because of how flexible and easy it is. R is my baby, and I am still learning to be ok at it. It's quirky as hell and requires some getting used to. Because it is so extensible, learning all the different packages can also be a pain in the dick. At this point, I am learning about basic stuff that allows you to produce your own shit, i.e. S3 and R6 and object-oriented programming. The amount of free content on the web makes learning this stuff solely based on your time available and willingness to learn. Did that come even remotely close to answering your question?

[wizzy]

Can also contribute. Analytic consultant focused on, well, analytics and data science at a systems integrator. Also a law school grad but transitioned away after no-offered from biglaw summer program.

How long did it take you to learn R, Python, SQL, SAS, etc.? Which did you learn first?

SQL is stupid easy. It took me maybe two months to be good enough to query out my info *reliably*. I focused on R because I am not a software engineer and will never put shit into production. Most of my data analysis is ad hoc, so R made sense. If I had to do it all over again, I might have done Python simply because of how flexible and easy it is. R is my baby, and I am still learning to be ok at it. It's quirky as hell and requires some getting used to. Because it is so extensible, learning all the different packages can also be a pain in the dick. At this point, I am learning about basic stuff that allows you to produce your own shit, i.e. S3 and S6 and object-oriented programming. The amount of free content on the web makes learning this stuff solely based on your time available and willingness to learn. Did that come even remotely close to answering your question?

Yup. I've dabbled in SQL and really rudimentary R but haven't learned Python. A lot of people say that Python is the best beginner language to learn for SWE purposes, but I'm assuming that if you're going for big data, you need R or Python but preferably both eventually?

If you were trying to get into analytics/data science and learn languages, where would you start in terms of putting together a learning/study plan (just in terms of which sites/resources/content to go through)?

[Kali]

Can also contribute. Analytic consultant focused on, well, analytics and data science at a systems integrator. Also a law school grad but transitioned away after no-offered from biglaw summer program.

How long did it take you to learn R, Python, SQL, SAS, etc.? Which did you learn first?

SQL is stupid easy. It took me maybe two months to be good enough to query out my info *reliably*. I focused on R because I am not a software engineer and will never put shit into production. Most of my data analysis is ad hoc, so R made sense. If I had to do it all over again, I might have done Python simply because of how flexible and easy it is. R is my baby, and I am still learning to be ok at it. It's quirky as hell and requires some getting used to. Because it is so extensible, learning all the different packages can also be a pain in the dick. At this point, I am learning about basic stuff that allows you to produce your own shit, i.e. S3 and S6 and object-oriented programming. The amount of free content on the web makes learning this stuff solely based on your time available and willingness to learn. Did that come even remotely close to answering your question?

Yup. I've dabbled in SQL and really rudimentary R but haven't learned Python. A lot of people say that Python is the best beginner language to learn for SWE purposes, but I'm assuming that if you're going for big data, you need R or Python but preferably both eventually?

If you were trying to get into analytics/data science and learn languages, where would you start in terms of putting together a learning/study plan (just in terms of which sites/resources/content to go through)?

If you are going big data, you will need spark and scala. But very, very few companies actually have big data. But yeah, R and Python are both good at certain things. I'd recommend Python before R as Python continues to grow in popularity while R still has its niches (ad-hoc, biopharma, reserach and academia.)

As far as recommended study courses for data science: start with the basics. Datacamp is by far the most beginner friendly site I've come across for this shit. They teach you, step by step, each part of the programming side and will get you introduced to the basics on stats (if you aren't already familiar). Once you get that done, I highly recommend https://www.coursera.org/learn/machine-learning Andrew Ng's machine learning course on coursera first, and then move to Mostafa's course https://work.caltech.edu/telecourse.html for a more in depth look at shit (this one was challenging for me). Just be prepared: Ng's course is in matlab, which sucks ass. But the point is understanding how the models work, not necessarily the tool you use to implement them.

[suralin]

yeah i work almost entirely w ~big data~ here and i and most of our data scientists use both python and R, depending on the potential analysis

personally i love python and still suck at R, but the latter is undoubtedly powerful af for really complex data analysis. we have a number of econ phds in data science and they produce scarily good / incomprehensible stuff using R for ecosystem and incrementality takes

if you do go the python route, highly recommend jupyter notebooks

re: big data, it's mostly a fad and kali is v right that most companies don't need it at all. you can scale sql way farther than you'd think. if you do need to get into hadoop, it's literally just learning hive which is basically sql

same thing tbh with ML. i have a GBDT model in prod rn but literally the hardest thing about it was feature engineering. for the vast majority of use cases, linear regression is good enough. for other cases, there's so many plug-and-play options now

but playing around w deep learning is fun (get into deepfakes on the ground floor!!)

[suralin]

Yup. I've dabbled in SQL and really rudimentary R but haven't learned Python. A lot of people say that Python is the best beginner language to learn for SWE purposes, but I'm assuming that if you're going for big data, you need R or Python but preferably both eventually?

If you were trying to get into analytics/data science and learn languages, where would you start in terms of putting together a learning/study plan (just in terms of which sites/resources/content to go through)?

sorry that didn't really answer your question

there are so many resources now, the hard part is choice overload: https://medium.freecodecamp.org/i-ranke ... 5dc7e3eb8e

would pick based on how hands-on and project-oriented it is

[wizzy]

Thanks for the recs, Kali and suralin. I've been meaning to start learning this stuff but keep putting it off. Definitely gonna check out the links.

[wizzy]

What's the typical career progression like for a software engineer (and for a data scientist)? I'm kinda familiar with the CS UG -> SWE I -> SWE II -> SWE III route generally, but what are the different skills you have to learn at each stage, and then where do you go from there?

[suralin]

What's the typical career progression like for a software engineer (and for a data scientist)? I'm kinda familiar with the CS UG -> SWE I -> SWE II -> SWE III route generally, but what are the different skills you have to learn at each stage, and then where do you go from there?

it depends a lot on the company, but big tech seems to have coalesced around a similar system (see levels.fyi)

at some of these companies IC (individual contributor) <-> manager tracks are parallel and pay equally, so you can be an IC and not manage anybody ever. we have IC SWEs at FB at the director+ level making millions/yr

re: different skills, tbh, technical skills are often just table stakes at a big 4-like company. the difference between an ic3 and an ic5 is only minimally that an ic5 can churn out more code and knows more things. mostly the difference is in increased scope, noticing opportunities, cross-functional influence, not just good code but good code that delivers impact (e.g., leads to product direction changes or metric movements), and effecting multiplicative impact (e.g., mentoring). there's ofc the fabled 10x engineer but those are rare. basically aspie coders don't fare well

edit: oh also you can't really talk about career progression w/o talking about job hopping. SV's turnover rate is insane. the easiest way to get promoted is to grind interview prep, interview at a peer company, and get upleveled when hired. this makes for an even bigger incentive when you consider most RSU initial grants end at 4 years and negotiating for more comp is so much easier when you're changing companies

[wizzy]

What's the typical career progression like for a software engineer (and for a data scientist)? I'm kinda familiar with the CS UG -> SWE I -> SWE II -> SWE III route generally, but what are the different skills you have to learn at each stage, and then where do you go from there?

it depends a lot on the company, but big tech seems to have coalesced around a similar system (see levels.fyi)

at some of these companies IC (individual contributor) <-> manager tracks are parallel and pay equally, so you can be an IC and not manage anybody ever. we have IC SWEs at FB at the director+ level making millions/yr

re: different skills, tbh, technical skills are often just table stakes at a big 4-like company. the difference between an ic3 and an ic5 is only minimally that an ic5 can churn out more code and knows more things. mostly the difference is in increased scope, noticing opportunities, cross-functional influence, not just good code but good code that delivers impact (e.g., leads to product direction changes or metric movements), and effecting multiplicative impact (e.g., mentoring). there's ofc the fabled 10x engineer but those are rare. basically aspie coders don't fare well

edit: oh also you can't really talk about career progression w/o talking about job hopping. SV's turnover rate is insane. the easiest way to get promoted is to grind interview prep, interview at a peer company, and get upleveled when hired. this makes for an even bigger incentive when you consider most RSU initial grants end at 4 years and negotiating for more comp is so much easier when you're changing companies

Oh man, I hate the nebulous delivering impact buzzwords, but I'm guessing it's different on the technical side when you have tangible code that you've written and software development metrics you can be assessed on.

How much of grinding interview prep is "just do leetcode bro?" What else goes into the process of prepping for a CS interview?

And then in terms of getting in the door initially, what's the baseline set of credentials you need to be considered for a software engineer role? I know that you always hear about the coding bro who just has a high school diploma, but I'm assuming that's not your typical big 4/unicorn SWE hire.

[suralin]

yea it can definitely still be nebulous — whenever i see or hear impact now i automatically append ~~

leetcode is like 80% of it. the rest is system design (primer), data or front end knowledge depending on the type of role, and being a normal human

yeah the industry is more meritocratic than most, but the whole idea you can bootstrap from HS diploma to big tech is pretty unlikely. baseline set of credentials is likely a quantitative-y UG major and a couple coding projects you can point to. many coding bootcamps are also p legit. but tbf it will always be an uphill climb when you're competing w CS majors and entry-level roles don't have low supply rn

[Kali]

What's the typical career progression like for a software engineer (and for a data scientist)? I'm kinda familiar with the CS UG -> SWE I -> SWE II -> SWE III route generally, but what are the different skills you have to learn at each stage, and then where do you go from there?

Depends on the data scientist. Most places have "Junior Data Scientist" positions that they are starting to open up. I'm not good enough to be called one (I think of myself as closer to an analyst than a scientist at this point). Also, and Sur let me know what you think, it seems that Data Scientist is typically a title reserved for Phd's that they guard like fucking lunatics and everyone else is referred to as an analyst.

[suralin]

What's the typical career progression like for a software engineer (and for a data scientist)? I'm kinda familiar with the CS UG -> SWE I -> SWE II -> SWE III route generally, but what are the different skills you have to learn at each stage, and then where do you go from there?

Depends on the data scientist. Most places have "Junior Data Scientist" positions that they are starting to open up. I'm not good enough to be called one (I think of myself as closer to an analyst than a scientist at this point). Also, and Sur let me know what you think, it seems that Data Scientist is typically a title reserved for Phd's that they guard like fucking lunatics and everyone else is referred to as an analyst.

lol yup that is definitely a thing

e: there's some thrash generally w defining what it means to be an analyst vs a data scientist vs a data engineer vs a SWE. data roles are relatively new

[Kali]

Yeah. I have engineers as closer to software engineers, i.e. putting data models and shit into production.
Analysts do the data cleaning and statistical testing.
Scientists do what analysts do as well as model selection, validation and work with engineers and business leaders to make the work actually matter.

[wizzy]

I've heard about Amazon doing weird things with stock vesting. Facebook, Google, Microsoft, and other tech companies seem to do an even vesting schedule with their initial stock grant over 4 years (25/25/25/25), while Amazon backloads so you get screwed if you leave early (5/15/40/40). I've also heard that Amazon adjusts your comp at the end of each year to account for the increase in the value of your AMZN stock (e.g., if your target comp is $200k, they'll make adjustments to your base/bonus so your all-in comp will be exactly $200k, while FB/GOOG pay you your base/bonus independent of how much your stock has appreciated, so if your $300k in RSUs increase from $75k/year to $150k/year, that's just more money in your pocket).

Do you know if that's accurate info?

[wizzy]

Another question: FB perks. Are they basically like Google with the free gourmet food, snacks, on-site gym, laundry and dry cleaning, barber's shop, arcades, sleeping pods, etc.?

[suralin]

I've heard about Amazon doing weird things with stock vesting. Facebook, Google, Microsoft, and other tech companies seem to do an even vesting schedule with their initial stock grant over 4 years (25/25/25/25), while Amazon backloads so you get screwed if you leave early (5/15/40/40). I've also heard that Amazon adjusts your comp at the end of each year to account for the increase in the value of your AMZN stock (e.g., if your target comp is $200k, they'll make adjustments to your base/bonus so your all-in comp will be exactly $200k, while FB/GOOG pay you your base/bonus independent of how much your stock has appreciated, so if your $300k in RSUs increase from $75k/year to $150k/year, that's just more money in your pocket).

Do you know if that's accurate info?

yup that's accurate though FB has actually recently eliminated the initial grant's 1 year cliff (for hires starting in 2018 and on). there was also never a cliff for yearly RSU refreshers; it and the bonus both get a multiplier depending on perf rating. and yeah FB definitely doesn't do that adjustment thing, that would blow (I got my initial grant when stock was ~$145 so )

[suralin]

Another question: FB perks. Are they basically like Google with the free gourmet food, snacks, on-site gym, laundry and dry cleaning, barber's shop, arcades, sleeping pods, etc.?

yea pretty much. at orientation they told us there's 50+ restaurants on campus—i'm partial to the bbq place and pho/ramen shop myself. re: transportation there's the wifi shuttles, valet (though this isn't much of a perk, more of a requirement bc we're growing so fast), and on-demand car service (a la uber). vacation is p generous too at 21 days from the start + unlimited sick leave. but the main thing to get used to was how often we have happy hours; there's literally 2-3/week w sushi and an open bar etc

[kandra]

Any thoughts on pivoting from CS to law? I am a CS undergrad with programming internship experience and experience with spark, cpp, golang, sql etc. After my internships and based on the advice of friends/family, I figured that programming isn't for me and law, my original passion, is where I want to be.

I guess I'm asking if you have any thoughts on keeping up my cs skill bank throughout law school and using it in the future?