It's hard to know in another domain (ie low-level stuff) what projects are bite-sized enough to achieve but also are challenging enough to allow one to level-up.
I totally identify. Of course, I'll second Deron's recommendation of HMH and especially Casey's intro to C.
In terms of bite-size, I have been slowly making my way through DanZaidan's Making a game in C from scratch
series. It's a great complement to HMH. HMH goes super in-depth on making an ambitious, professional 3D game, which gives it the virtue of covering many aspects of game dev in depth. However, it's a much bigger project than someone like me would choose for their first low-level programming project. By contrast, Dan's series tackles a smaller-scope game while still explaining everything along the way, so it's great for a beginner like me to see the entire process more quickly.
Outside of game dev, I've also thought for a while of making some web-dev-related things that I wish existed but are against the grain of the industry. For example, I've cried many tears working with CMS frameworks that have very counter-productive ways of letting you reuse their code. I would like something more based around the principles outlined in Casey's Designing and Evaluating Reusable Components
lecture (basically, functions I can call to do what I want, rather than a monolithic engine that I write hooks for).
I've heard that the software running this very site is sort of along those lines. It's sort of a stripped-down Python/Django setup, but without too many of the heavy abstractions from Django. I'd be interested in the source code but it's not public AFAIK.
Finally, I find this new Büchse project
very interesting. A similar idea occurred to me but I don't know enough about network programming, threading, etc. to tackle it as well as Stas is doing.