Machine learning is one area where it takes a bit more imagination to do it in Ruby, but you can easily wrap both compiled code and Python for that matter and still do your own work in Ruby if that's what you want to do.

I'd strongly argue that irrespective of your preferred language (ok, almost, before someone brings up Intercal or something) if you in any domain think it doesn't support what you want to do, what it actually means is that you don't understand how to achieve it with that tool. That's fine, and maybe it's not the right tool for you, but it's not the fault of your tools

And if you're not happy with what you're working on, you're unlikely to stick with it long enough to have any impact. And even you do, if the result is being miserable it's not worth it.