Author · New Book

AI Agents in Production

Infrastructure Engineering for Reliable LLM Systems

AI agents are easy to demo and hard to run. This book is about the gap in between — the infrastructure, patterns, and engineering discipline that take an agent from a flaky prototype to a reliable, observable production system.

• Taming chaos: error handling, retries, and graceful degradation
• Orchestrating tools, APIs, and multi-step agent workflows
• Guardrails against hallucination and non-deterministic failure
• Observability, evaluation, and cost control at scale
• Orderly, repeatable deployment of LLM systems