The talk is particularly useful for developers of Python applications which does heavy computation, with or without NumPy - for data science, research etc. Python extensions allows for creation of high-performance applications, which can compete with C or C++ based ones. There are more than one framework to achieve this (for example, Cython, Pythran and Numba) with similar syntaxes but different underlying implementations. The talk surveys the state of the art of creating extensions and introduces Transonic (https://transonic.readthedocs.io). Transonic is a pure-Python package acting as a unifying front-end for writing extensions with the aim to enhance the developer experience.