Recently quit my job at Facebook and will be joining Google soon. Got a lot of good information from the Blind community with regards to interview process, compensation, expectations, etc. and wanted to pay that forward. E5@FB (ML profile, no masters/PhD, worked there for 6 years) Base: 190k Unvested equity: 400k Interviewed for L6@Google 2 coding rounds 2 ML design rounds 1 Googliness and Leadership (behavioral) Pretty sure coding and behavioral went great. ML interviews went good-ish. I was optimistically expecting an L6 offer. Got an offer for ML Engineer, L5@Google (couldn’t negotiate level) Base: 198k Bonus: 15% = ~30k Equity: 558k -> negotiated up to 775k, in return for cancelling loops with Amazon/Twitter Sign on: ~30k Humble tips: - Interviews are not crazy hard, even if you don’t get nowhere in first half of the interview, don’t worry about failing — that way, you might have a chance to get to the right solution in final 15 mins (it’s not easy I know. meditation helps) - Be honest, in interviews and in negotiation. It does require you to really know yourself, your motivations for leaving, what you’re seeking in a new job, priorities, long term goals — these are good things to know (again, meditation helps); but will make your life easier by smoothing communication + negotiation. - Try to never give out your current TC or numeric expectations. If you do, set high expectations, like 40-50% more than what you want. - Don’t lie your numbers. If you do, firstly, inform yourself of levels and pay ranges and secondly, be consistent, write them down, the process might go on for 30-40 days and good recruiters write down *everything* that you might have said to them, *ever*. - You don’t always need competing offers to negotiate. There are many other kinds of leverages — in my case: (a) my option that I might just stay put at FB (b) enthusiasm of the destination team’s manager (c) the fact that I was considering cancelling interviewing at other places for higher offer. Happy to answer more questions, and to hear feedback on what I could have done better.
Congratulations! How were the ML rounds? What level of detail/math was expected in those rounds?
Hard to generalize based on 2 interviews. Ultimately discussion revolves around designing some solution and there is some leeway to steer your discussion. Math and detailed knowledge of specific techniques comes in handy when you have to trade off computational complexity vs effectiveness of different methods (for example - finding a good sequence of objects by statistical methods that consider all permutations vs RNN models), justifying what loss function to use, estimate data size based on model parameters, etc. A decent chunk of discussion is also around general ML knowledge like dataset specs, when to use what type of techniques, training procedures, designing multiple stages to help with scale, pros and cons of different metrics, etc. I’d say fairly equal focus on both of these aspects.
Thanks, that's great info. Could you share how you prepared for ML design? Would greatly appreciate it.
You say that "interviews aren't hard" but there are many aspects to why an interview could be hard. For some it's LC problems. For others it's communication and clearly explaining their thinking, or behavioral rounds. Which were the most difficult and why? While prepping did you: Do mock interviews? If so, how many? Do any LC questions or similar practice?
Excellent point. The last thing I want to do is to undermine the difficulty of interviews, but I don’t consider myself a genius and want people to really believe it’s doable (hence my use of phrase 'interviews aren’t *crazy* hard'). I found ML design interviews harder than rest, because there are many pitfalls: it’s super open ended, easy to digress (especially when you and the interviewer have different pictures in mind and you’re going deep technically). If you wait to check with the interviewer whether you’re going in the right direction it could be read as lacking independence. I did no mocks (I was too overloaded with work tbh). I treated interviews as regular closed room meetings with a manager or colleague - I think that imagination took some pressure off me. I did ~10 mostly medium LC questions, before the interview. Some more details about this is in the first comment thread.
How did you get slotted for L6 interviews?
I’ve been E5 for almost 4 years almost, and have been expecting promo for a while now and was a little annoyed by the delay (maybe justified, I have a lot to learn). Anyway a recruiter for Google reached out to me 2 years ago for L5, and I didn’t want to move then. This time I reached out to him and asked for an L6 interview and he helped. In short, I asked (and got lucky).
@OP what's your yoe? what did you do before fb?
Why not wait for E6 promo? L5-6 at Google is very hard.
how hard?
I wanted to quit my FB team. So my promo efforts would get nullified anyway. 5->6 is hard in FB as well, so I might be up against similar odds. Another big factor for me is a manager that can give me thoughtful feedback, and that’s also helpful in getting a promo. I talked to my prospective manager at Google three times, and I got a feeling that he has a good understanding of the path form L5-L6. He has plans to create an open L6 position sometime end of next year, which further corroborates that. So I decided to trust that.
Negotiated just for cancelling amazon interview? How did that work
Hard to say what it looks like from the company’s POV. I think they were trying to reduce the possibility of non conversion of the offer after investing so much into the process. There could be other factors like maybe team wanted to fill a position quickly, or recruiter wanted to close the deal, or something else. But its now obvious to me that there are many factors at play that could be used as leverage for a price that the offer committee feels good about.
This worked for me as well when I joined fb. The recruiter increased my TC if I agreed to cancel linkedin and twitter interview. So I did.
whyd you leave fb
Left because (a) got the team I desired at Google, great product, great place for ML learning (b) I was at FB for 6 years, and had stopped growing, in part, due to incompatibility with management style, especially post covid.
So you have 6 YOE?
Yes
You mentioned in the other thread that you are in E5 lvl for 4 yrs so yo u got to E5 in 2 yrs. That seems impossible.
Congrats OP. Could they verify if you were interviewing with amazon and Twitter? Wondering Why would a company negotiate so much just because you say you’re interviewing elsewhere.
I don’t think they could verify. My best guess is that as long as they pay to reduce chances of no-conversion having invested significantly in the interview process, hiring committee, etc.
This offer is lower than fb e5 offer of 205/700/100. Will you stay if fb matches it
Mostly no. I’m not optimizing for money only. Been in FB for 6 years learnt a lot, not anymore.
How much lcs have you done? Did you move because you love google as a product / what’s the reason ?
What’s lcs? Moved because (a) got the team I desired at Google, great product, great place for ML learning (b) I was at FB for long time, and had stopped growing, in part, due to incompatibility with management style, especially post covid.
Lcs=leetcode right?