Principal AI Engineer Handbook
A reference handbook for engineers building production AI systems, covering reference architectures, runnable labs and interview material for a set of recurring AI-infrastructure problems.
The handbook is organized into a Learn section of modules, a Build section of production-quality labs, reference architectures for recurring AI-infrastructure problems, interview questions placed next to the material that answers them, one-page reference lookups, decision records explaining why the platform is built the way it is, and printable cheat sheets, with the four main sections cross-linked so a topic can be read from idea to code or back again.
As described, it currently has sixteen Learn modules, eleven labs each paired with a matching architecture page, three interview tracks, thirteen reference lookups and five cheat sheets; every lab passes ruff, mypy in strict mode and its own test suite in continuous integration, and every page also has a two-voice audio episode with its transcript, synthesized locally. It is built as a site run locally with pnpm and deployed to Cloudflare, and is licensed MIT.