Articles liés à Building Production AI Agents with AWS Bedrock AgentCore

Building Production AI Agents with AWS Bedrock AgentCore - Couverture souple

Saxena, Saurabh

 
9798187090723: Building Production AI Agents with AWS Bedrock AgentCore

Synopsis

Your working demo just broke in front of real users. Now what?

Every developer who has built an impressive AI agent proof-of-concept eventually hits the same wall: getting it to actually work in production. Reliability, security, cost control, memory, observability, evaluation — the demo never prepared you for any of it.

This book is the bridge from POC to production.

Using AWS Bedrock AgentCore, you'll build one complete, real-world project — an invoice-processing agent called InvoiceBot — from the very first line of code to a fully deployed, enterprise-grade system. No toy examples that fall apart under load. One serious agent, built the right way, chapter by chapter.

What you'll learn:

  • Set up your environment and ship a working agent in 30 minutes
  • Host agents on AgentCore Runtime and connect real tools through Gateway
  • Give your agent memory, secure it with Identity, and ground it in your own documents with RAG
  • Run code safely, add observability, and see exactly what your agent does
  • Defend against prompt injection and other agent-specific security threats
  • Evaluate whether your agent actually works — golden datasets, LLM-as-judge, and continuous evaluation
  • Optimize cost and token usage, handle errors, test, and set up CI/CD
  • Master agentic workflow patterns, anti-patterns to avoid, and responsible-AI governance
  • Learn from real production incidents — and the concrete solutions that fixed them

Who this book is for:
Whether you're a Python developer, a "vibe coder" who learns by building, a solutions architect, or a technical founder, this book meets you where you are. No prior experience with AI agents or distributed systems required — every concept is explained with plain-language analogies before any code appears.

Bonus: A complete interview-prep chapter with 70+ scenario-based questions and answers for the Forward-Deployed Engineer role — covering technical architecture, consulting, stakeholder management, behavioral questions, and Amazon Leadership Principles.

By the end, you won't just understand how production AI agents work — you'll have built one.

One project. Every service. Production ready.

Les informations fournies dans la section « Synopsis » peuvent faire référence à une autre édition de ce titre.

Vos droits au titre de la garantie légale : Chaque consommateur de l'UE bénéficie d'au moins deux ans de protection au titre de la garantie légale pour tout bien de consommation vendu dans l'UE. Voir plus