What it does This repository houses supporting materials for AI Engineering (O’Reilly, 2025): chapter summaries, study notes, curated resources, prompt examples, case studies, and a grab-bag of tools including a conversation heatmap generator for ChatGPT and Claude exports. The book itself is a 150,000-word framework for adapting foundation models—LLMs and LMMs—to real applications, with an explicit warning that it is “NOT a tutorial book” and contains minimal code.

The interesting bit The angle is deliberately anti-hype: Huyen focuses on fundamentals over tools because “tools become outdated quickly, but fundamentals should last longer.” The repo also positions the book as a companion to her earlier Designing Machine Learning Systems, mapping where traditional ML engineering ends and AI engineering begins—a boundary many teams are currently stumbling across.

Key highlights

• Covers end-to-end decisions: whether to build, how to evaluate, when to finetune vs. not, and how to make models “faster, cheaper, and secure”
• Addresses specific production pathologies: hallucination detection/mitigation, RAG strategies, agent evaluation, feedback loops
• Case studies drawn from actual projects, reviewed by a bench of practitioners including a ChatGPT co-creator and former Fortune 500 CIOs
• Explicitly warns before deep technical sections—useful for mixed-audience teams
• Repo is actively being populated; author notes “more resources in the next few weeks”

Caveats

• The repository itself is currently sparse; most value lives in the book or linked external resources
• One “fun tool” (the heatmap notebook) is more of a curiosity than a utility
• No code-heavy tutorials here; developers looking for copy-paste implementations will need to look elsewhere

Verdict Worth bookmarking if you’re an AI/ML engineer, engineering manager, or technical PM trying to move from prototype to production. Skip it if you want hands-on frameworks or API-specific walkthroughs—you’ll find neither.