Using Amazon Bedrock : learn to architect, secure and optimize generative AI applications on AWS / Renaldi Gondosubroto.

Format:

Book

Author/Creator:

Gondosubroto, Renaldi, author.

Series:

Tech today.

Tech today

Language:

English

Subjects (All):

Amazon Web Services (Firm).

Artificial intelligence.

Web applications--Development.

Web applications.

Physical Description:

1 online resource (483 pages).

Edition:

1st ed.

Place of Publication:

Newark : John Wiley & Sons, Incorporated, [2026]

Summary:

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.

Contents:

Introduction

What Does This Book Cover?

Who Should Read This Book

Part 1 Introduction to Your New Generative AI Playground

Chapter 1 Introduction to Generative AI on AWS

The Current Generative AI Landscape

Introduction to Foundational Models

Core Technologies

Key Attributes

Transfer Learning

Generalization and Adaptation

Creativity

Emergent Behavior

Interoperability

Efficiency Challenges

Recent Advances in Generative AI

Industries Using Solutions with Amazon Bedrock

Healthcare

Retail

Finance

Advertising

Tasks Performed by Generative AI

Summarization

Software Development

Question Answering and Knowledge Retrieval

Chatbot Enhancement

Generative Design

Ethics and Security

Ethics

Security

Technical Challenges

How Amazon Bedrock Approaches AI Challenges

Fairness

Explainability

Privacy and Security

Robustness

Governance

Transparency

Life Cycle of a Generative AI Solution

Problem Definition and Scope

Data Collection and Preparation

Model Selection and Tuning

Training and Evaluation

Integration and Deployment

Monitoring, Security, and Compliance

Feedback and Improvement

Tech Setup for Using Generative AI on AWS

Python for Developing Generative AI Applications

Understanding How Bedrock Works

Understanding Pricing in Bedrock

Amazon Bedrock Capabilities

Forms of Generation

Text

Images

Embeddings

A One-Stop Solution

Amazon Nova

DeepSeek

Claude

Stable Diffusion

Llama

Command and Embed

Mistral

Creating Knowledge Bases

Fine-Tuning

Agents

Safeguards

Spaces for Experimentation

Amazon Bedrock Playground

PartyRock.

Working with Other AWS Services

Unified Data Access

Security and Authorization

Monitoring and Logging

AI Assistants

A Customer Service Chatbot Solution with Bedrock

Summary

Chapter 2 Prompt Engineering with Foundational Models on AWS

Crafting Good Prompts

Best Practices in Crafting Good Prompts

Clarity and Specificity

Context and Background

Iteration and Feedback

Break the Tasks into Steps

Testing and Experimentation

Advanced Techniques

Chain of Thought Reasoning

Meta Prompting

Templating

Using the Model Context Protocol for Dynamic Prompting

Tackling Malintent in Prompts

Prompt Injection

Jailbreak Attempts

Hallucinations

Evaluating Prompt Performance

Learning About Zero-, One- and Few-Shot Inference

Zero-Shot Inference

One-Shot Inference

Few-Shot Inference

Configuring Your Bedrock Environment

Working with Parameters for Foundational Models

Controlling Randomness and Diversity of Responses

Temperature

Top P (Nucleus Sampling)

Top K

Controlling Length

Max Tokens

Stop Tokens

Repetition Penalty

Cost Optimization Based on Prompting

Experimenting with Playgrounds Provided by AWS

PartyRock

Building Your First PartyRock App

Configuring Your Generated PartyRock App

Practical Exercise: Finding the Right Prompt

Part 2 Core Generative AI on AWS

Chapter 3 Building Applications with the Amazon Bedrock API

Overview of Using the Amazon Bedrock API

API Endpoints for Using Bedrock

Invoking the Model

Using Textual Models Through the API

Configuring the Appropriate Parameters

Inference Parameters

Troubleshooting the Parameters

Formatting the Right API Calls

Streaming Responses

Handling Exceptions

Prompt Caching for Cost and Latency Optimization.

Cost Optimization and Architectural Best Practices

Using Stable Diffusion for Image Generation

Using the Bedrock Image Playground

Coding Text-to-Image Implementation

Maintaining Context in Generative AI Chatbots

Understanding the Requirements of a Generative AI Web App

Architecting the Application Concept and Functionality

Choose Your Technology Stack

Set Up the Backend

Integrate the AI Model

Develop the Front-End

Handle the User Input and Display AI Responses

Test and Iterative Improvements for Model Performance

Deploy

Building Your First Generative AI Flask Application

Chapter 4 Working with Multimodal Foundational Models

Multimodal Foundational Models

Creating Prompts for Multimodal Foundational Models

Text and Multimodal Foundational Models: Key Differences

Integration of Multiple Inputs

Rich Contextual Information

Visual Detail Specificity

Dealing with Ambiguity

Iterative Feedback

Best Practices for Working with Multimodal Foundational Models

Well-Crafted Prompts

Balanced Detail

Cultural and Societal Sensitivity

Syntactic Consistency

Iterative Testing and Refinement

Leveraging Model Strengths

Encouraging Creativity and Diversity

Applying Model Architecture and Training Data in Prompts

Enhancing Context with Multimodal Foundational Models

Image-to-Image Transformation

Image Inpainting

Image Outpainting

Visual Question Answering

Image Captioning

Working with Amazon SageMaker JumpStart

Preparation

Configuration of Permissions and Variables

Model Retrieval and Endpoint Deployment

Endpoint Interaction and Response Handling

Practical Exercise: Creating a Movie Recognizer

Data

Search Index

Retrieving Results

Demo Application

Summary.

Chapter 5 Fine-Tuning Foundational Models on AWS

Fine-Tuning Foundational Models

Hyperparameters for Fine-Tuning

Epoch

Batch Size

Learning Rate

Learning Rate Warmup Steps

Putting This into Practice

Instruction-Based Fine-Tuning on AWS

Creating a Dataset for Fine-Tuning from Your Output

Formatting of a Multitask Instruction Dataset

Converting a Dataset to Be Used for Instruction

Fine-Tuning with Instruction on Bedrock

Setting Up for Fine-Tuning

Installing Libraries Necessary to Perform Fine-Tuning

Importing Libraries and Initializing Variables

Creating an IAM Role and Policy for Using Fine-Tuning

Writing the Instruction and Iterating Over Data Points

Defining the data_transform Function and Processing Data Partitions

Converting and Uploading the Datasets for Use in Fine-Tuning

Creating the Fine-Tuning Model Based on Llama 3.3

Using Tokenization

Using fmeval

Fetching Variables from Before and Importing Relevant Libraries

Creating the Fine-Tuning Job

Creating Provisioned Model Throughput

Monitoring the Status of the Deployment of the Model

Invoking the Fine-Tuning Model

Cleaning Up the Resources

Fine-Tuning on Amazon SageMaker JumpStart

Amazon SageMaker JumpStart

Deploying a Llama Model with Amazon SageMaker

Fine-Tuning Using SageMaker Studio

Using SageMaker JumpStart in SageMaker Studio to Fine-Tune

Cleaning Up Endpoints

Advanced Considerations for Fine-Tuning

Chapter 6 Performing Retrieval-Augmented Generation on AWS

Retrieval-Augmented Generation

In-Depth Explanation of RAG

Broad Spectrum of RAG Applications

Advantages and Ethical Considerations

RAG Compared with Other AI Models

Practical Limitations of RAG

Fundamentals of RAG on AWS

Process of Retrieval-Augmented Generation.

Selecting an Appropriate Data Source

Selecting Data

Define the Scope and Requirements

Assess Data Relevance and Diversity

Ensure Data Quality

Consider Data Security and Compliance

Use of Metadata and Annotations

Testing and Iteration

Selecting a Vector Database

Role of Vector Databases in RAG

Practical Application

Selecting the Appropriate Vector Databases

Refining Embeddings for RAG Use in Vaccine Information Retrieval

Selecting an Embeddings Model

Configuring LangChain

Preparing Your Data

Question Answering

Using Agents for Working with Foundational Models

Working with Agents for Amazon Bedrock

Uploading Datasets to Amazon S3

Creating a Knowledge Base

Creating an Agent from the AWS Web Interface

Setting Up a Question-Answering Task

Chapter 7 Optimizing Performance for Foundational Models

The Challenges of Compute and Memory with LLMs

Memory Management

Compute Resources

Inference Optimization

Evaluation and Refining Performance of Foundational Models

Reinforcement Learning

Approaches to Reinforcement Learning for Book Reviews

Learning Loop for Book Reviews in Bedrock

Automatic Model Evaluations

Human Worker-Based Model Evaluation Jobs

Model Evaluation Tasks

Text Generation

Question and Answer

Classification

Creating the Appropriate Prompt Datasets

Built-in Prompt Datasets

Custom Prompt Datasets

Understanding Model Evaluation Job Results

Automated Reports

Human Report Cards

Distributed Computing Approaches to Achieving Optimization

Embracing Data Parallelism

Advancing with Model Parallelism

Harnessing Hybrid Parallelism

Leveraging Elastic Scaling

Implementing Effective Load Balancing

Ensuring Fault Tolerance and Recovery

Using Prompt Caching.

Optimizing Performance with Step Functions and Lambda.

Notes:

Description based on publisher supplied metadata and other sources.

ISBN:

1394382634

9781394382637

1394406703

9781394406708

OCLC:

1547123958