The AI Engineer's Library
Production handbooks for engineers shipping AI.
Two books that cover the full operational stack for LLM features in production — from tracing and evals to multi-step agents and recovery. Written for backend and platform engineers.
Paperback · Available now on Amazon
Book 1
Observability for LLM Applications
Tracing, Evals, and Shipping AI You Can Trust
The production handbook for backend and platform engineers shipping LLM features. Covers OpenTelemetry GenAI semantic conventions, the 2026 tool landscape (Langfuse, LangSmith, Arize Phoenix, Braintrust), evals as a first-class observability signal, cost tracking, drift detection, and incident response. 18 chapters, ~88K words.
What you'll learn
Available now in Paperback and Hardcover. Kindle eBook coming soon.
Book 2
Agents in Production
Building, Tracing, and Shipping Multi-Step AI You Can Trust
How to build, trace, evaluate, guard, deploy, and recover LLM agents in production. Covers LangGraph, OpenAI Agents SDK, Anthropic Claude Agent SDK, CrewAI, Microsoft agent-framework, Meta's Agents Rule of Two, and the full agent operations stack. 18 chapters, ~94K words.
What you'll learn
Available now in Paperback and Hardcover. Kindle eBook coming soon.
The series
Two books. One operational stack.
Book 1 builds the observability foundation — tracing, evals, cost tracking, drift detection. Book 2 applies it to agents — multi-step workflows, guardrails, deployment, recovery. Read Book 1 first for the foundation, then Book 2 for agents specifically.
Observability
18 chapters · ~88K words
The foundation. Instrument, trace, evaluate, and monitor LLM features with OpenTelemetry and the modern tool landscape. Know when your AI is drifting before your users do.
Agents
18 chapters · ~94K words
The application. Build, trace, guard, deploy, and recover multi-step LLM agents using every major framework. From single-tool calls to autonomous workflows.
Written for
Ship AI you can actually trust.
About the author
Gabriel Anhaia
Senior Software Engineer based in Berlin. 10+ years of experience building scalable backend systems for fintechs and high-growth companies.
Also the author of the Actually Learn series, Hexagonal Architecture in Go, The Complete Guide to Go Programming, and creator of Hermes IDE.
Ship AI with confidence
Stop guessing.
Observe. Build. Ship.
Two books covering the full production stack for LLM features and AI agents. Available now on Amazon.
Paperback · Available now