Ask HN: Minimum viable transition to AI?

Briefly, I am a mid-30s full-stack (or generalist) developer of no particular pedigree (college dropout, mostly self taught, small time consulting, and working for small technology startups).

I was 'good' at math as a kid but neglected it as I was older and, as a result, have always been intimidated by ml/ai. Until recently, I had it in my head I missed the boat on having the math foundation necessary to do ai.

On a whim (while on sabbatical), I took Geoffrey Hinton's Neural Network course. This required significant tangential studying to get my linear algebra and calculus up to snuff. All in all, I spent about a month where most of my time was either spent on Hinton's course or the math foundations necessary to understand it.

Feeling very excited, I was looking for what to study next, assuming I could probably get myself hire-able within a year or so--in the meantime, I would consult in my "old" work, and build a couple portfolio pieces (my first is just about done already).

Searching the internet and asking around in AI chatrooms, I came across the fast.ai courses. A few weeks after that (this week), I have a tenuous grasp of how to use pytorch, especially with the fast.ai library, and I am able to get shockingly decent results on my own applications of the techniques.

Now I wonder if there is enough demand for deep learning practitioners that I might be able to continue my learning as a junior in a lab somewhere more or less immediately.

My end goal is to build robots that will facilitate mass automation of human labor. And, to do it in a way in which the technology/IP behind the automation will be public/free of patents and restrictions.

Salary/location not particularly important.

Given all of the above, what does the minimum viable CV look like to start getting income in this space? Is there room for non-seniors to grow into capable researchers?

Thanks for any insight.