Books

Bestselling book on building LLMs. A practical, project-driven manual for engineers who want to understand how modern language models are built — and where they fail — by writing every layer themselves. From a scalar autograd engine to RLHF to fused specialists, in 36 hands-on projects with deliberate sabotage experiments. Build it. Break it. Measure it.

A best-selling book. The practitioner's guide to Claude Code in production. Thirty-one chapters covering the agent loop, tools, hooks, MCP, the Claude Agent SDK, permissions, multi-agent orchestration, evals, observability, and cost engineering. Includes a full walkthrough of Anthropic's financial services reference agents. Code from real production systems, not toy examples.

The most important impact of AI on software development isn't that it writes code faster — it's that it changes what you can delegate. Agentic Programming is a practical roadmap for climbing the AI Fluency Ladder: from prompting to agentic workflows, verified execution, and ultimately autonomous software development.

Master language models through mathematics, illustrations, and code―and build your own from scratch!

With the advent of LLMs and AI coding assistants, we are generating code at an ever-accelerating rate. How do we manage this influx of code without getting fatigued and overwhelmed? How do we ensure coding standards are maintained? What practices and tools can we put in place to amplify our efforts while not introducing more bugs?

Skip the black-box frameworks. Build a production-grade AI coding agent from scratch in pure Python - cloud or local, tested with pytest, all in a single file.

A clear, illustrated guide to large language models, covering key concepts and practical applications. Ideal for projects, interviews, or personal learning.

A hands-on guide to designing, building, testing, and deploying secure stdio and SSE MCP servers in Python and TypeScript (447 manuscript pages).

The fastest practical path to understanding harness-driven development as a complete system. Through clear visual diagrams and a real repository mapped to the book, you will see how a harness guides AI agents, evaluates their work, detects drift, enforces constraints, supports repair, and keeps software evolution visible, verifiable, and under control.

A scenario-based study guide for the Claude Certified Architect Foundations (CCAR-F) exam. Learn to reason like an architect across all five domains, then prove it on two full practice exams.

The definitive guide to agentic software engineering with Codex CLI, from prompting and AGENTS.md fundamentals to multi-agent orchestration, CI/CD integration, security hardening, and enterprise deployment across 28 hands-on chapters.

You don’t need massive compute or big-tech resources to build real AI. This hands-on guide shows you how to build and fine-tune your own small language model—from scratch—using accessible tools like Google Colab. Learn transformers step-by-step, train and align models on your own data, and deploy practical AI systems that run on consumer GPUs. If you’ve ever wondered “Can I build my own model?”—this book proves the answer is yes.

Vector stores don't think — they search. They find fragments that sound like your query, then forget they ever looked. Every session starts from nothing. Every context window is a memory that dissolves at sunset.But the deeper problem isn't amnesia. It's that when agents do remember, they remember in someone else's house — on servers you don't control, in formats you can't inspect, under terms you didn't write.Memory Graph is a book about building something different: persistent, structured, queryable memory that lives inside your application — no external servers, no data leaving your process, no infrastructure you don't own. An embedded graph database that travels with your agent the way a nervous system travels with a body.You'll learn how to model not just facts, but relationships between facts. Causality. Temporal ordering. The layered structure of meaning that makes memory more than a search index. You'll build ontologies that enforce what can be known and how. You'll combine graph traversal with semantic search — so your agents find not just what's similar, but what's connected.The result is an agent that remembers the way you do: structurally, contextually, privately — with memory that belongs to you.

Metagraphs for Agentic AI: Beyond Triples, Beyond HypergraphsFrom Knowledge Graphs to Knowledge ArchitecturesThe triple is not enough.Every AI engineer building agent memory hits the same wall. You model a meeting as a knowledge graph triple — and immediately lose the fact that five people were in the room, a decision was made, and that decision caused three downstream actions. You reify. You flatten. You create workarounds. And your "knowledge graph" becomes a tangle of auxiliary nodes that machines can traverse but no human can reason about.This book shows you the way out. What You'll Learn Metagraphs are graph structures where edges connect sets of nodes to sets of nodes — and where edges themselves can be referenced as first-class nodes. They are the missing data structure for AI agents that need to remember, reason, and coordinate like humans do.This book takes you on a complete journey:Hypergraphs first. You'll learn what they are, why they matter, and where they break down. You'll implement them three ways — in SQL, in LadybugDB (Cypher), and in TypeDB — so you understand the tradeoffs viscerally, not just theoretically.Then metagraphs. You'll see how metagraphs solve the fundamental hypergraph problem (edges that can't be nodes), explore RDF named graphs as a lightweight metagraph, and implement full metagraphs in the same three database paradigms with production-ready, commented code.Then the big ideas. Semantic Spacetime. Holonic systems. Human cognitive architecture mapped to graph structures. Multi-agent coordination. Promise Theory for autonomous AI networks. This is where metagraphs stop being a data structure and become an architecture for intelligence. Who This Book Is For You're a software engineer, AI researcher, or knowledge graph practitioner who builds real systems. You've used Neo4j or RDF stores. You've built RAG pipelines. You've felt the limits. You want to know what comes next.No PhD required. Every concept comes with working code in SQL, Cypher, TypeQL, SPARQL, and Python. What Makes This Book Different This isn't a theoretical monograph. It's the distillation of two and a half years of research, 130+ published articles, and hands-on implementation at the intersection of knowledge graphs and agentic AI.Every chapter bridges theory and practice. You'll read about Basu and Blanning's formal metagraph definition — and then build the schema in PostgreSQL. You'll learn Mark Burgess's Promise Theory — and then model a multi-agent coordination protocol as a six-layer promise graph. You'll understand why labeled property graphs are secretly metagraphs — and what that means for your Neo4j deployment today. 18 Chapters. Three Parts. One Argument. Part I — The Hypergraph Foundation (7 chapters): From the knowledge representation crisis through hypergraph theory to three complete database implementations.Part II — The Metagraph Solution (5 chapters): Metagraphs as the answer, RDF named graphs as a bridge, and three full metagraph implementations with detailed commentary.Part III — Theory Meets Practice (6 chapters): Semantic Spacetime, labeled property graphs, AI memory and human cognition, holonic systems, agent-to-agent interaction, and Promise Graphs for network-of-networks coordination. The Core Thesis If you want AI agents that reason like humans, you need knowledge structures that capture how humans actually organize knowledge — not as flat collections of facts, but as nested, hierarchical, context-rich, temporally-aware structures where relationships themselves carry meaning and can be the subject of further reasoning.Metagraphs are that structure. This book shows you why, and how to build with them.

Everything you really need to know in Machine Learning in a hundred pages.