Uber: Marketplace (SF and PA) Google Brain: Tensorflow for Mobile/Embedded devices. (Mountain View) Current TC: 165K YOE: 2 Worked on ML projects and have a Master's in ML. Base/RSU/Yearly Bonus/Sign-On Uber: L4: 150K/330K/15-30%/25K Google: L3: 130K/Don't know/15-30%/20K Google will try matching. Both seem like cool opportunities. Uber Pros: 1. I know the team well and like the work they do (building productionizable ML models on cool datasets) 2. Fast paced, I'll have more ownership. It's like working in regular software role. (Agile and more organized workflows) 3. Good money, IPO coming up! Uber Cons: 1. Not Google 2. SF (I prefer the boring South Bay) 3. IPO coming up! Google Pros: 1. Google Brain 2. Experienced ppl in the team (Just 3 but staff or very experienced) 3. Will do ML work. 4. Team is 40 percent formed. (Joining in early stages) Google Cons: 1. Not the typical kaggle/modelling. More focused on running TF and optimizing ML models to make them work on different smaller devices (infra + ML). 2. Intimidating. I'm literally listing this out to clear my head. Tough choice to make. @google and @google brain folks, your thoughts on SWE-ML work? ease of switching? @uber folks, how is work at marketplace? How is it being an ML engineer at Uber? @rest, your thoughts on salary and teams? Answer the poll too!
Money speaks.. ask for enough stock.
You won't be doing anything fancy as an L3, doesn't matter what they say.
IPO is this week so google should ideally match your Uber offer..
This poll didn't help at all..I'm still confused. 🤦♀️
Brain is a highly recognized. L3 TC can go 250k+. Check comp fyi. Embedded/Mobile optimization for TF/TF Lite would help you gain specialized subject matter. Google optimizes not for just Pixel, but for Edge use cases. Depending on your preference, you may or may not enjoy it. But I’ve seen this type of expertise extremely valued at today’s hot ML Infra market; where expertise in Deep learning framework, compiler or accelerator (GPU, TPU, Apple Neural Engine, Qualcomm SNPE etc) are super scarce. Self-driving industry also looks for this expertise. I know folks with ~2 yoe post-PhD getting 400k+ offer.
The problem is I'm used to the more data sciency role and worry that working on this might take me on another path. But it is Google brain, I'll be focussed on modelling ML algos so they run on these devices. So hopefully less of the actual infra work. Confused if I should take a leap of faith into this unexplored domain. Or stick to the usual data sciency role cause I want to know more of that as well.
Like I mentioned separately, you will look better as a janitor, or whatever else, at Google Brain than you probably would at Uber.
What did you work on at Bloomberg?
Structured data, traditional ML algos for personalization.
Google sounds like infra development. I don’t think you’ll enjoy it if you’re into building models