Maybe you should look into some other kind of programming hobby? Any why can't you do ML on your own anyway? Does it require too many computational resources? (serious question, I know almost nothing about the field)
Personally, I could easily work on some kind of OSS project if I suddenly found myself with lots of free time. I can do C, C++, Python, etc., so it wouldn't be hard for me to find an existing project to contribute to, or make my own.
I think I was unclear. We're in agreement. I already do plenty of it in my spare time, like you suggest, and I'd love to expand the spare time stuff. What I was trying to add was that since I'd prefer to do it less than 40 hours per week, I wish there was a paid way to do it less than 40 hours per week. The status quo doesn't have much room for paid part-time work in this line of business, and that makes me sad.
And depending what you're doing, it can require ridiculous computational resources (like hundreds of thousands of dollars to train a model), but that's far from the norm. If you aren't trying to advance a couple particular states of the art, you can get by with a fancy computer and patience, or a pretty reasonable AWS/GCP bill, probably comparable to any other fairly expensive hobby.
I agree about the part-time thing too; considering how much of a skills shortage the industry is constantly crying about, and how well-paid developers are (which means that, if they live frugally and don't have kids, they don't really need to work to retirement age if they don't want to), it really should be possible to have part-time positions for people who want to stay current and work on interesting problems but not have to dedicate themselves to it full-time, and not be stuck only doing OSS stuff (which is great, of course, but not really doable with some fields).
Personally, I could easily work on some kind of OSS project if I suddenly found myself with lots of free time. I can do C, C++, Python, etc., so it wouldn't be hard for me to find an existing project to contribute to, or make my own.