It’s expensive and time consuming to change, tune, and retrain an LLM to modify an agent’s behavior. Self-Improving Agents: How to engineer adaptive agent harnesses shows you how to design AI agents that measurably get better in production without fine-tuning, weight updates, or waiting for the next model release. Written by Micheal Lanham, author of AI Agents in Action, this hands-on book equips you with techniques to transform your agent’s harness—context, memory, metacognition, tools, and code—into a measured, versioned, auditable improvement loop that adapts and improves as it runs.
Micheal Lanham
Getting an agent to complete a task is one challenge. Making it improve over time—and proving that a change actually helped—is another.
This book explores how agents can improve without retraining the underlying model. The focus is on the agent’s harness: its context, memory, tools, and code, with changes measured, versioned, and audited rather than accepted on faith.
The running project is HelixAgent, a RAG-based agent that gains a new self-improvement layer as the book progresses. Along the way, the book tackles practical questions about trustworthy evaluation, learning from live traffic, and keeping changes under control with human approval gates and rollout safeguards.
It’s aimed at intermediate Python developers building or operating LLM agents. The book is now available through Manning’s Early Access Program, with four chapters ready to read and more to come.
- Full details: Self-Improving Agents - Micheal Lanham
Don’t forget you can get 45% off with your Devtalk discount! Just use the coupon code “devtalk.com” at checkout 