about this book
This book is a practical guide to building an AI agent from the ground up. Its goal is not to teach any single framework but to give readers a complete mental model of what an agent is made of by implementing each layer themselves. By the end, readers should be able to open the source code of any popular agent framework and follow along, as well as design and operate an agent tailored to their own needs.
Who should read this book
This book was written with the following readers in mind:
- Developers who have used LLM APIs but remain unclear about the internals of an agent
- Engineers who want to see what lies behind the magic of frameworks
- Teams who need to build agents tailored to their own data, tools, and workflows
- Practitioners who must operate agents reliably in production
What this book asks of you is a grasp of Python basics and some experience calling APIs. A deep background in machine learning or natural-language processing isn’t required. What is required is the willingness to read, run, and modify the code yourself as you learn. These are the only prerequisites for this book.
How this book is organized: A road map
The book is divided into two parts that span 10 chapters. Part 1 builds the smallest working agent from the ground up. Part 2 grows that agent into a practical system and then shows how to evaluate, operate, and improve it over time. Each chapter extends the code of the preceding one, so the agent isn’t rewritten along the way—only extended.