Available now in early release
Building AI Agent Platforms
Tools and Tactics to Scale AI Applications Across the Enterprise
Building a single AI-powered application is relatively simple. Creating a platform that lets many teams build, deploy, and operate agents securely and reliably is the harder and more valuable problem.
This book is a practical roadmap through that entire lifecycle: platform strategy, architecture, model access, runtime, observability, experimentation, governance, and reliability.
Read on O’ReillyInside the book
From first capability to production platform.
Strategy
Choose the capabilities worth investing in and build a platform teams actually want to use.
Architecture
Design model access, runtimes, orchestration, and developer workflows for scale.
Evaluation
Measure agent quality, learn from production signals, and improve through experimentation.
Reliability
Build in observability, security, guardrails, and operational discipline from the start.