Skip to content
eBook Charts
CA
AI Agent Design Patterns

AI Agent Design Patterns

Ganapathy Lakshmanaperumal

$0.99

Buy on Amazon View on Apple Books

Prefer it in print? See every format on Amazon →

Price on Apple Books. Amazon’s price may differ.

Read a lot? Try Kindle+ * · Compare subscriptions

Published
Genres
Computers & InternetSoftware

Description

An independent, practical catalog of design patterns for AI agents, written for software architects, senior engineers, AI engineers and computer science students. Every pattern uses one template: intent, problem, forces, solution, structure, a vendor-neutral Python sketch, trade-offs, failure modes and when not to use it. It starts with agent anatomy, workflows versus agents and a reference architecture. It then catalogs workflow patterns, prompt and context patterns, reasoning and planning patterns such as ReAct, plan-and-execute, reflection and tree search, tool-use, memory and retrieval patterns, and multi-agent orchestration: supervisor, hierarchical teams, handoff, blackboard, debate and swarm. Later parts cover human-in-the-loop patterns, reliability patterns such as checkpoints, idempotency keys, sagas and circuit breakers, evaluation and guardrail patterns, security patterns such as least privilege, dual LLM and quarantine, cost and latency patterns such as routing, caching and cascades, and interoperability through the Model Context Protocol and agent-to-agent protocols. One fictional insurance claims platform runs through every chapter. Anti-patterns, a selection guide, a hands-on lab, review questions with answers and a glossary complete the book. This book is an independent publication and is not authorized, sponsored or endorsed by any company whose products it mentions. All trademarks belong to their respective owners and are used only for identification. This book was drafted with generative AI tools and checked with automated tools. Tools, features and prices change often; confirm them in current documentation.