The Python that ML code is actually written in: types, generators, decorators, context managers and a test suite.
10h · 7 lessons · 1 challenge
Not an introduction to programming. This is the subset of Python that ML codebases lean on — iterators and generators for data loading, decorators for instrumentation, context managers for devices and precision, dataclasses for configuration — plus enough packaging and testing that your take-home runs on the reviewer's machine.
Skip it if you already write Python professionally; the placement quiz will tell you.
A slow data script and a profiler. Find the hot line, fix it, run again. Scored on the speedup you get in five attempts.
Worth 350 XP.
