Using Amazon Bedrock

A from-scratch roadmap to building generative AI solutions on AWS with Amazon Bedrock

In Using Amazon Bedrock: Learn to Architect, Secure and Optimize Generative AI Applications on AWS, accomplished Software Engineer, developer advocate, and AWS Community Builder, Renaldi Gondosubroto, delivers an in-depth walkthrough of Amazon Bedrock, the keystone generative AI service on the Amazon Web Services cloud. Gondosubroto offers a start-to-finish guide of the service and its capabilities, from prompt engineering with foundational models to building applications using the API, working with multimodal models, and fine-tuning.

This book provides hands-on instruction on Amazon Bedrock from an experienced developer and AI specialist. It’s packed with real-world code samples, proven best practices, and techniques that result in reliable, secure, and cost-effective generative AI solutions. You’ll also find:

• Demonstrations of robust governance guardrails and strategies for reducing development time
• Repeatable, industry-grade frameworks for continuously iterating and optimizing models in production
• Contemporary updates that incorporate the latest announcements made by Amazon at the re:Invent conference in December 2024 and beyond

Perfect for cloud architects, artificial intelligence engineers, and software engineers, Using Amazon Bedrock is an insightful, original, and practical roadmap to generative AI on AWS that explains the AI fundamentals you need to understand to get started in AWS generative AI development and the hands-on techniques you’ll use every day to transform those concepts into efficient, working solutions.

Introduction xxi

Part I Introduction to Your New Generative AI Playground 1

Chapter 1 Introduction to Generative AI on AWS 3

Chapter 2 Prompt Engineering with Foundational Models on AWS 43

Part II Core Generative AI on AWS 87

Chapter 3 Building Applications with the Amazon Bedrock API 89

Chapter 4 Working with Multimodal Foundational Models 133

Chapter 5 Fine-Tuning Foundational Models on AWS 183

Chapter 6 Performing Retrieval-Augmented Generation on AWS 225

Chapter 7 Optimizing Performance for Foundational Models 265

Part III Advancing the Building Blocks of Generative AI 309

Chapter 8 Security and Privacy for Deploying Generative

AI Architectures on AWS 311

Chapter 9 Building End-to-End Applications with Generative AI 363

Chapter 10 Sustainability and Scalability with Amazon Bedrock 405

Appendix A Configuring Your AWS Account 433

Appendix B Installing Python, Jupyter Notebook, and LangChain 439

Index 443

RENALDI GONDOSUBROTO is an AWS Community Builder, an accomplished software engineer with over a decade of architecting solutions on AWS, and a developer advocate in the tech community. He currently holds all 14 AWS certifications. He has extensive experience developing AI solutions for small and large enterprises.

A start-to-finish playbook for building generative AI solutions on AWS with Amazon Bedrock

In Using Amazon Bedrock: Learn to Architect, Secure and Optimize Generative AI Applications on AWS, AWS Community Builder and artificial intelligence expert Renaldi Gondosubroto delivers a hands-on walkthrough of Amazon Bedrock, Amazon Web Services’ central, fully managed service for building generative-AI apps. The author dives deep into the techniques you can use to create generative AI solutions, including prompt engineering with foundational models, guidance for working with multimodal models, fine-tuning, and interacting with the API.

You’ll find optimized code samples, best practices, and step-by-step guides for creating reliable, secure, and cost-effective generative AI tools. You’ll also discover repeatable frameworks for iterating and optimizing production models and explanations of the latest capabilities including guardrails and utilizing Claude 4.

Using Amazon Bedrock is a must-read guide for cloud architects, AI engineers, developers, and programmers interested in building efficient, working solutions using generative AI technology. It covers every useful feature offered in Bedrock and explains exactly how to use them as you develop sustainable and scalable software that solves real-world problems.