This repo at a glanceFrom the README

📥 Download PDF / EPUB (Recommended) — The PDF / EPUB is recommended for offline reading, with the best typesetting; you can also read online (supports language switching, chapter folding, highlights and notes, and is automatically updated with every push). Agent = LLM + Context + Tools — Built around this core formula, the book uses 10 chapters to take AI Agents from principles to engineering practice. The full text, figures, and all 109 accompanying experiments are open source; you're welcome to run through the experiments yourself. 📚 The companion volume, Deeply Understanding AI Infra: Quantitative Analysis and System Design, is now open source and available to read: github.com/bojieli/ai-infra-book To develop good model-based applications, you also need to understand the infrastructure these applications run on. The companion volume covers the AI Infra that supports model training and inference: where parameters and context state are stored, how computation is executed, and how multiple accelerators work together. 📢 Version 2.0 changes (compared with 1.4): The book manuscript in this repository has been upgraded from version 1.4 to 2.0.