Dear admins (if any), I've encountered the subject. Please give me some hints how to speed up my Python solution, because I've tried all the performance tricks I know and I still get TL2.

You can post some details about your solution, for example the complexity your program has. A solution with N^3 complexity should fit in time (with c++).
Also post the core function (loops) of your code, maybe you have some mistakes and enter some endless loop.