Hey everyone, just for a little background on me. I will be graduating from a good school in March with my BA/MS in CS. I have done a series of internships for startups in ML and SWE for amazon. My goal in the longterm is to work as a research engineer in a top ML lab. Unfortunately, I only started doing research in my last year of school. I won best senior thesis and will have a second author paper at a good ML conference. I havent gotten any interviews for research scientist/engineer positions, and I was only able to procure two job offers post graduation. Capital one is a machine learning role, but I was wondering if the name recognition of Amazon would be of more assistance for a good ML role in the future. Amazon is seattle C1 is Virginia. C1 is MLE rotational program and amazon is SWE in aws simulation. Again, I am optimizing for a career in ML Amazon is 185 tc normal breakdown y1 and 175 after C1 is 155 tc y1 and 132 tc after
Amazon is better than capital one any given day. If you want a career as a research scientist, higher education (MS, PhD) is must
I will be graduating with an MS and some research experience and still doing some research on my own time once I start working. I love the space and am determined to transition to a role in 18-24 months. I am just wondering if you think amazon SDE will open more doors than C1 MLE for roles?
C1 is crap (any decent recruiter knows that) HMs expect that a person who has worked at a decently sized big firm such as Amazon, can figure out things on their own I got my PhD in aerospace but recently interviewed for Apple Silicon (HW)
You can pivot to another role after you’re in, but it won’t be easy. I’m just kind of wondering why you applied for a swe job in the first place. Being a junior SWE in any big tech company is probably not going to be the experience you’re hoping for if your ultimate goal is research. You’ll have limited scope and many of the problems you’re actually qualified to solve might be pawned off to dedicated science teams and you’ll effectively be the infra guy
My ultimate goal is not research, its research engineering. I like the interesting technical problems surrounding ML/AI (data compression, GPU optimization, scaling, online differential privacy, etc.) The skills involved in SWE seem pretty applicable here. Also, I'm not the most competitive application for research engineer roles given my background (I applied to every new grad job I found with a role like this, but didn't get any bites).
Again, i don’t think those are the types of problems you will work on as a junior engineer, or even midlevel engineer. A quicker path to that outcome is likely to pursue a research-oriented work. I’m not saying it won’t happen, but there are a lot more teams that don’t fit that path than the ones that do in big companies, and mobility is very low these days given the market so it’s not like you can jump around like you used to to find that fit
The moment you join amazon... just look for ML team to transfer to
How feasible is this? I am assuming everyone is gunning for the ML/AI roles for internal transfers. Should I try to start networking with people through slack right away, or attend in person events or something?
Go to internal job page, and message hiring managers that have ml projects. DO NOT STAY ON A TEAM that you see no future esp for new grad. Dont waste more than 3 months if the team they assign you is a dumpster fire. And no. ML is not as hot as you think it is. Amazon actuslly does have many projects that are interesting.
Amazon for sure!
Do you have any insight into what the AI projects are like at capital one?
More than that please read about their performance management, really toxic culture
I am unsure what you mean by research engineer? If you mean, research scientist, although many will say you don’t need it, but you need a PhD. Master’s thesis is not enough to get research scientist positions. ML engineer is a slightly different in that you try to productionize models to run at scale (which includes tuning, ML ops etc.) Research scientist performs research and publishes journal articles. These positions are highly competitive and you will need to have published pedigree to be considered for this role in FAANGs.
Yes ML engineer is the role I am gunning for
Amazon does not have ML engineer job title because they see engineers as fungible. There is applied scientist (which is in between research work and ML engineering work) and research scientist. If you are joining as a SWE at amazon, be prepared to be evaluated as an engineer during performance review.
Performance management at C1 is AIDS. I’d rather deal with Amazon’s pip quotas while earning higher comp and having better name recognition. Not to mention the comp increases from L4->L5->L6->etc. are much more substantial. We have managers at C1 earning under $200k.
Damn I used to make more >200 as new grad at msft
Confirm. I have been at C1 for 5 years and still have no idea what is happening with performance management. Complete mess.
My two cents. I can't talk about C1, but have some experience in Amazon. Even when switching roles is on the papers, in the practice is not a straightforward process, specially on the current situation. Having said that, if you decide to continue with Amazon, you have to be lucky with the team your assigned to. There are teams when you can be exposed to ml things, and others that are way apart. After some time in the company I think your best ally is your manager. As long as you are clear and transparent about your goals your manager can help you with that (also you have to be lucky with that, not everyone is open to let their people move that easily, but for switching roles AFAIK your manager should support you). I personally think that switching roles will take at least a couple of years (I may be wrong for sure, but even for switching teams in same role, without an exception, requires a tenure of 1 year). My recommendation, if you don't have other offers, would be to accept Amazon but keep looking for other opportunities. For sure having it on your resume can help you pass some first stages on other hiring processes.
Switching roles within a team at Amazon takes many years from what I've seen
Go to grad school.
Go to Amazon, apply for ML jobs or a PHD
Congratulations on the offers, OP! They’re pretty much the same company nowadays, so you’re better off going for Amazon and bouncing to another FAANG company in a year. Ideally you would keep looking for another company since Amazon has a bad reputation, but with the new grad market being in shambles, it’s best to take the chance at Amazon and use it to get your career off to a decent start. I would caution you to be wary of which org you’re joining; Amazon has had numerous layoffs, re-orgs and PIPs throughout the year, and Music was recently affected by this. Make sure that you know exactly what you’re getting into and document everything. Good luck! TC: 105K
Thanks for the reply! You dont think being out of the ML space for a year would hurt my chances of working in ML again?
No worries. You can easily rebound to ML provided you have the right research and experience under your belt. Depending on how the market goes, you may be able to switch over to more ML heavy down the line. Amazon has its flaws, but from what I’ve read they provide a decent amount of opportunities to learn and grow—if you take advantage of that and continue your ML research, you’ll be fine. Congratulations once again!