21.3Building Agentic Systems

A practical guide to AI system engineering: the agent loop, tools and structured outputs, state and memory, orchestration across many agents, human-in-the-loop design, reliability, observability, cost engineering, and security. Written by someone who tests software for a living — evaluation and testing of agentic systems is a first-class chapter, not an afterthought.

• On Leanpub: leanpub.com/building-agentic-systems