How to learn agentic AI, end to end.

A working engineer's guide to building agentic systems — not chatbots that answer questions, but software that plans, calls tools, verifies its own work, and ships to real users. The book starts from how LLMs actually work, walks you through writing an agent loop from scratch in 4 chapters of dual-API code (Anthropic and OpenAI), then takes you through the parts most material on this topic skips: production observability, cost discipline, safety, evaluation, the four distinct agent shapes (code, computer use, research, multi-agent), and finally how to stay current as the field moves. Opinionated where the field has converged, honest where it hasn't. Built by someone learning it — for anyone learning it.