Series reference

#The AI Engineering Textbook

A companion to the Deep Dives. Where each dive's README is a lab manual that tells you what to run, the TEXTBOOK.md in each repo is the lecture: the what, when, where, why, and how behind the code. This page introduces the textbook and links every chapter.

#Why this exists

If you have ever taken a science or engineering course, you know the two halves of it. There is the lab, where you build things with your hands and watch them work, and there is the lecture and its textbook, where someone explains what you just saw and why it behaves the way it does. You need both. The lab without the theory is a sequence of steps you followed without understanding; the theory without the lab is trivia you cannot use.

This series has always been strong on the lab. Every dive is a repository you walk through, every concept is a small runnable program, and every section ends with something to run. What the series lacked was the lecture, the connected story that explains why each piece exists, where it came from, and how it fits with everything else. The theory was there, but scattered across README sections, code comments, and the moments of insight you only got if you happened to read closely.

The textbook is that missing half, gathered and told as a story. Each chapter takes one dive and explains its subject the way a good teacher would: starting from the problem the technique was invented to solve, tracing a little of its history, building the mental model you need, and being honest about the tradeoffs and the places it breaks. There is very little math. There is a lot of "here is why this works, and here is the moment the industry figured it out." You can read a chapter before doing its lab as a preview, or after as a way of connecting what you built to the larger picture. Either order works; doing both is the point.

#How to read it

The chapters follow the same sequence as the series. The eight core chapters build on each other, each adding a layer until you are operating a real application end to end. The bonus chapters branch off where they are most useful and can be read whenever their subject comes up. If you are new, read in order. If you know what you need, jump to it; each chapter stands on its own while linking to its neighbors.

If the question "what is a language model, really?" is still fuzzy, start with the primer in HOW-LLMS-WORK.md, which sits underneath the whole textbook. The GLOSSARY.md defines every term the chapters assume, and MODELS.md covers the specific models and their prices.

#The core path

The eight chapters that build on each other, in order. The thread runs: build the call, ask it well, ground it in knowledge, measure it, let it act, harden it, and operate it.

#The bonus chapters

Standalone chapters that extend the core path. Each notes where it slots in.

#A note on how these were written

The textbook holds itself to the same standard as the labs it accompanies, the one written down in AUTHORING-LESSONS.md: the reader believes what is on the page, so what is on the page has to be worth believing. The chapters teach the honest tradeoff over the tidy-but-false claim. They tell you where techniques fail, not just where they shine. When they cite a number or a result, it is one the labs actually produced, including the surprising and the unflattering ones. The goal is not to make the material sound impressive. It is to make it clear, accurate, and, where possible, enjoyable to read.

Back to the series overview · The decision guide: CHOOSING.md · The vocabulary: GLOSSARY.md