You might see whether your local community college has good classes that supplement what you already know, and what companies hire software interns/co-ops/newgrads from there. (Long ago, this, and luck, was how I jumped from self-taught, to a co-op student at a first-rate engineering company, and maybe it can still work.)
And some colleges/universities have options for students outside of a Bachelor's degree program. These might or might not teach better than a community college. (But "Harvard Extension School", for example, will probably get past more resume screeners than a community college.) Both might teach you more than a bootcamp. And, again, look for something set up for internships, or with relationships with specific employers.
Beware that the field is absolutely flooded by everyone in their dog, and it's hard to distinguish yourself. I suppose the upside is that there's so much not-so-stellar work being done, that a newbie could their foot in the door at a typical place, survive while they contribute, and grow from there.
But this flood also means that people hiring will reach for a kind of gatekeeping (to narrow down the pool of candidates, and to try to cement themselves as a superior in-group), which, at the moment, means you probably need to practice for Leetcode hazings.
If you have a lot of time, doing open source is an option to eventually stand out, but if you do this, pick something in a space where people are hiring. ("I wrote a JIT compiler for a dynamic dialect of Haskell that targets GPU" wouldn't even get you considered for 99% of jobs, and most of the remaining 1% would insist you regurgitate Leetcode algorithm as performance art so they can "verify that you know algorithms" or so that they "can see how you think". "Yes, I looked at your GitHub, and it didn't have many stars." :)
Also keep in mind that Copilot-like tools integrated into IDEs are going to let a lot of low-skilled developers launder open source code, to pass off as their own (justifying it as "I let the tool do the boilerplate, so I can focus on the more difficult, higher-level problems"). I don't know how the near-term evolution of this will play out, but it might be best as a force multiplier for low-skilled, low-quality work. You might want to be thinking about how to be on a path to eventually stand out for high-quality work, for which companies won't trust this generation of automation (even if they were willing to gamble with a Copilot-induced GPL lawsuit, and VC due diligence doesn't start smacking down for this).