You can either leave or find a way to make your job fun. I've repeated "Be the change you want to see in the world" about a thousand times since I've started me

You can either leave or find a way to make your job fun. I've repeated "Be the change you want to see in the world" about a thousand times since I've started mentoring people, and it's really the only thing you as an engineer can do. Their architecture is a clusterfuck? Come up with ways to improve it and present to some slice of the company that will listen (e.g. director of engineering, C-levels if the company is small enough). Ultimately, you can make almost any job tolerable just by putting in the legwork. If you're not comfortable doing that or you feel like there's nothing you can do, then it's time to leave. That's how I decided to leave my last job, I realized I couldn't make any more change for the better and so it was time to leave and do something else. I will say atrocious code is one of the more fixable problems you can find yourself facing. It starts by being a great example and enforcing good practices on anything you get ownership of, but ultimately it will eventually involve a discussion with whoever's in charge to start setting larger organizational policy. You said it's an interesting niche, and that's an important consideration, but remember that very few companies are "unique"; someone else is probably doing this, and jumping ship is an acceptable thing to do in any job. You are far more valuable to the company than they are to you.