[piecerough]

As a FAANG employee, working with ML, what do you want to get from other companies, besides more money?

It's hard to have more chips, for example. You run less experiments, you have less throughput in an already computationally tight environment.

[angarg12]

Fully remote would be a pretty nice perk, ever since our companies forced us back to the office. I believe this alone might be one of the easiest ways for smaller companies to poach talent these days.

Other than that, I enjoy working in hard and impactful problems. I'm privileged that I was able to do that in FAANG, but it's unclear how long this will last. I know that is very trite to say that your company has changed, but I truly feel it. Our company missed the LLM train and we are scrambling to play catch up now. Also we mostly pivoted from bold and inspiring initiatives to efficiency and cost cutting. Most of the initiatives in our pipeline for this year are pretty dull "let's save some dollars here and there". Nothing wrong with that, but given the choice, the chance of working on a moonshot vs helping a trillion $ company save pennies might convince one to change jobs.

There is also the very trite but very true aspect of "small cog in a huge machine". I might engineer a new system that brings in tens of millions of dollars to the company, and I'd be lucky if 3 layers of management above me knows about it. Those kind of initiatives might influence the entire company in a smaller organization.

All in I recognize I'm quite privileged and doing pretty well in my current job, and that's why I'm passively looking. TBH one of the aspects is that I got FOMO from all the talk about ML engineers being in high demand. However my experience so far has been very underwhelming.

[AznHisoka]

Where have you heard that ML engineers are in high demand? I don’t think they are immune from the tech slowdown that’s affecting the industry. Of course, LLMs and AI are all the rage, but the demand for that is in the services/APIs not necessarily the demand for ML engineers

[lumost]

I suspect we'll see a broad slow down in classic ML approaches. If I just need a basic classifier -> GPT-4 can do an OK job. Spending an afternoon prompt tuning is vastly cheaper than a dedicated team of scientists and engineers curating datasets and running experiments. For many SMBs, this effectively eliminates the ML function.

[breather]

The traditional answer is stability. I think that's still true despite the sense of looming layoffs. Everyone always knew the layoffs would come one day anyway...

[cannonpalms]

It sounds like a Staff title may be in play, although it wasn't made clear whether this would be a promotion.