I did give Ruby an honest shot, the maturity of the ecosystem around numerical computation and machine learning (my field) isn nowhere near Python’s. That is to say, it’s not that Ruby was bad, but that Python was so far ahead of everything else that to deviate from that would mean spending an inordinate amount of time reinventing the wheel. I’ve certain fallen into this trap in the past (which caused to me extend my graduate school a year because I wanted to use niche tech which was cool but ended up having to spend time working around corner cases, which were solved problems) but sometimes going with the mainstream means you get a lot solutions for free and can focus on your problem.
Programming language choice is partly aesthetic —- some languages are more elegant than others —- but if you want to be employed and want to create real impact in the world, it has to be tempered with hard nosed pragmatism as well. You’re paid to deliver solutions to problems, not to optimize your preferences. Sometimes you can do both, but when you can’t, solving the problem must take precedence.
> I can honestly say I couldn’t have done the same with Ruby because the Ruby ecosystem and tooling didn’t support such impact size in my (and many similar) domains.
But did you give it an honest try, or is that an assumption? While no ecosystem I at python's level in the ML space, Ruby's contributions are not necessarily non-existent, and you'll find a few options when you actually look for it (sidenote: the pgvector extension for postgresql has been created by a prominent ruby ML developer)