Stay a Fullstack SWE or pivot into ML Infra?

Hi Blind, Currently have the following two offers (not naming names so I don't get doxxed), and still trying to negotiate up so nothing is set in stone, but the final numbers probably won't look too different: 1- Senior SWE at a late-stage SaaS startup that raised during the recent downturn, $310k ($195k Base/ $110k Equity / No bonuses but there will be equity refreshers), fully remote. Work will be building out a new product vertical. Pros: - Will be saving up a lot more from the base and geo-arbitrage (at least ~$45k annually, maybe up to ~70k if I actually try to save), and I could think of a 100 different things I can use that money towards (car, down payment, wedding, buying the dip) - More entrepreneurial and ownership driven culture - I have very strong belief that the product will be a hit and bring in millies in ARR, so lots of impact and potential for career growth within the company Cons: - Paper money, and their valuation is very, very high right now, so even in the best case scenario I don't expect to see the equity portion grow much. That said, their overall offering is super solid and I'm certain they can 10x in the next 4 years to justify their current valuation, so imo worst case scenario is a 20% loss there. - Almost no technical learning, the project is just a web app at the end of the day - WLB likely to suffer 2- ML Infra SWE (L4) at a FAANG's skunkworks division (i.e. Microsoft Garage), $260k (185/60/20 and no refreshers), hybrid in California. Work will be primarily productionalizing/scaling existing R&D pipelines and models, as well as building out the tooling to enable more rapid/reliable experimentation Pros: - Will definitely learn a lot, it's a problem space I'm not familiar with and my colleagues all seem very technical (everyone I interviewed with was at least a PhD) - The company and progression looks really good on my resume, and I've experienced first-hand how many doors that opens up Cons: - I don't know if I'll actually like the work, or if there's really a future demand for this type of role. Afraid of getting pigeonholed into DevOps stuff that I don't enjoy - I'm really not a fan of corporate bureaucracy and feel like there will be a good amount of that here Context about me: I love to travel around so the remote work is a huge plus, but I don't mind holding off on that for a bit if the opportunity makes sense, plus I'm sure I can work something out with my manager in the worst case if I don't actually need to be there. My long-term goal is to eventually start my own thing. I'm not the kind of person who's interested in the nitty-gritty details so maybe Infra isn't a good fit there, really aiming to be more of a generalist. At the same time, I feel like there's just too much technical learning to pass up on and I wouldn't be able to work on these types of "scale" problems or break into MLE later on in my career, while I could always come back to product at any time and still keep my skills sharp there by side-hustling. Would love more insight from people who've actually worked in an MLE role and can shed some light on the kinds of things I can learn and the career path it could open up. Current TC: 170k 🥜 YOE: 3 #machinelearningengineer #career #advice #offers