Usually if you use Cython, f2py, numba, or something similar, the goal isn't to create code that fits into the ecosystem of the compiled language - it's to optimize a kernel that fits into the Python ecosystem with minimal baggage.