Python Foundations

The first course of the Agentic AI Engineer path. This is not a syntax tour: every concept is taught by building a small primitive an agent is later assembled from, and every bug we drill is one that actually breaks agents in production.

You will work in pure standard-library Python and finish able to write the building blocks the rest of the path assumes: a message store that never corrupts shared state, a token-and-cost estimator, an input router, a tool registry, a typed tool function, a retry-with-backoff wrapper, and a config and eval-case loader.

Module 1 starts with the single hardest and most expensive idea to get wrong: how names, objects, and mutation really work, taught through the aliasing bug that silently corrupts an agent's conversation history.