AI Engineering Playbook
A structured, searchable learning path covering retrieval-augmented generation, embeddings, vector search, agents and Model Context Protocol, aimed at engineers who build with LLM APIs rather than train models.
The AI Engineering Playbook is an eleven-section learning path for applied AI engineering, built as a searchable site with dark mode, rendered diagrams, self-graded quizzes and progress tracking. Every page answers three questions in order: why the concept exists and what problem it solves, how it actually works one level down with the relevant mechanism and math, and what breaks in production at scale.
It is aimed at engineers building at the application layer who already know how to code and call an LLM API, and it explicitly excludes machine-learning training material such as PyTorch or backpropagation. Sections progress from LLM fundamentals through embeddings and search, vector database selection, and on to retrieval-augmented generation, agents and the production engineering questions around them, such as evaluating retrieval quality, when not to build an agent, and where latency and cost come from in an LLM pipeline.