If you think you really do like LevelDB try writing your own version of it from scratch. This will test your understanding of the fundamental principles. It will also test your motivation. Then try to explain to someone else how to do what you did.
Once you truly understand LevelDB then pick a thorny problem they are working on, propose a solution, gather feedback, and solve it. Wash, rinse, repeat. A couple years from now you could be the LevelDB expert you've always wanted to be.
Then a few years down the road maybe you'll find some key insight that everyone was missing and write your own database that solves that particular issue and you'll be the one that someone else is looking up to.