I have always preferred working on concrete problems rather than abstract theories.
The best way to learn mathematics is to do mathematics.