Most valuable skills for an early-stage startup engineer (new grad)

I'll assume you're mainly interested in tech skills. Soft skills obviously aren't specific to AI, although as the other commenters humorously point out, there are certainly differences between startups and big companies that you'll want to understand. One in particular is basically being comfortable with uncertainty and understanding how to navigate it. Good communication and emotional self-management are key here.

As far as tech skills, there isn't a one size fits all answer -- the landscape, even within "early-stage startups in the AI space", is simply too broad.

You can get a good idea of what tools are most commonly used in the area just by browsing the web, looking at blog posts and open source projects, etc.

More specifically, I recommend going to LinkedIn or whatever job posting site you prefer and finding a bunch (10-20?) job postings that really excite you and tallying up the most commonly required skills. Go learn those, build up a portfolio of online demos and git repos, and you'll probably be as prepared as you can be to navigate the job market.

Naivety of believing all the lies you’ll be told about your equity (if any!) in order to justify you working 80 hours a week for a shitty pay and massive tech debt, when you could just go work at Google (or similar) and be treated like royalty while learning how to write proper software.

Source: started my career in early stage startups, now at Google. I even got lucky to have made a couple bucks from my equity (so I am a massive outlier, statistically speaking), but it still wasn’t worth it.

I loved my first experience at an early-ish start (I was engineer 4), and it made me confident that I could create a working product with the knowledge I currently had.

My biggest negative on the engineering side was that we were almost always encouraged to neglect good, testable architecture in favor of faster releases. This lead to more production breakages and confusion about what is affected by what changes.

Since then, I've realized just how critical testing is to a system, and that it's not a "nice to have", it's a must, especially as your team doubles in size again and again.

I had a technical founder who didn't lead the engineering team always say "we need more tests" (unit specifically), while not allowing us to refactoring our codebase to actually make it testable (most of the API code was just regular PHP scripts that could span 1k+ lines of business logic, untested).

That's another thing: your technical founder of head of the engineering isn't always going to be the smartest person ever. I had that idea in my head that of course they'd be an incredible engineer since they're the one in charge, but that isn't always the case. There's nothing wrong with that either, because hopefully they're a strong leader and that matters more for their position.

I do want to say though, I've worked at other earlier stage startups that have much more attention put into design decisions and testing and it is much more my speed. I can't fathom not having an automated test suite that tests the majority of my codebase automatically to catch the little stuff that bites you later on.

If the business doesn’t have product-market fit, the single most valuable skill you can exercise wrt the company itself is zeroing in on that. Speed, business understanding, and speed again are the top 3 components of that

Whether or not that aligns with your own benefit is a different question entirely. If your comp is set up well and the company is supportive, it should be.

No family