Advanced Python

You can already write Python that works. This discipline is about Python that keeps working — after six months, in someone else's hands, under load, at three in the morning when the person on call is not you. That is a different skill from writing the feature, and it is mostly made of decisions you can name: where the contract lives, what the type checker is allowed to prove, which boundary gets a test double, whether that slow function is slow for the reason you assume.

The method throughout is to build the tool before using it. You write the raises context manager, the yield fixture and the parametrize decorator before you rely on pytest; you drive an event loop made of generators before you write async def; you build the task-graph engine behind a Makefile, the shrinker behind property-based testing, the registry behind metrics. This is not nostalgia for doing things the hard way — it is that a framework you have rebuilt stops being incantation, and you can predict what it does at the moment it surprises you. Production topics are taught local-first: in-process simulators for CI and services, with real registries, sockets and queues discussed as prose rather than pretended at. And the worked example is this repository — the make quality gate you are standing inside is the one the testing pillar dissects.