book

The Developer's Guide to AI

by Jacob Orshalick, Jerry M. Reghunadh, Danny Thompson

June 2026

Intermediate

320 pages

7h 45m

English

No Starch Press

Read now

Unlock full access

Cover Page
Title Page
Copyright Page
About the Authors
About the Technical Reviewer
BRIEF CONTENTS
CONTENTS IN DETAIL
ACKNOWLEDGMENTS
AUTHORS’ NOTE
PREFACE

INTRODUCTION
Generative AI and BeyondData Security and PrivacyTechnology ChoiceDownloading the Example CodeThe AI Coding Assistant Safety Recall NoticeWhat We’ll Cover
PART I: GETTING STARTED WITH AI
1. UNDERSTANDING LARGE LANGUAGE MODELS
Understanding Text GenerationAI vs. Traditional LogicThe Versatility of LLMsUnderstanding the LimitationsLimited Training KnowledgeHallucinationsModel BiasMulti-Step ReasoningImproving LLM ResponsesPrompt EngineeringContext EngineeringFine-TuningCustom ModelsBuilding Autonomous SystemsPlugging LLMs into Your CodeSummary
2. BUILDING YOUR FIRST LLM-POWERED APPLICATION
Calling an LLM by Using an APIVerify System RequirementsInstall Ollama and Run a Local LLMSend a Prompt to the Local LLMInstall Node.jsInitialize Your ProjectInstall Node.js PackagesCreate the Server FileLaunch Your Server and TestStreaming Your LLM ResponsesStream an LLM Response with OllamaStream a REST Response with ExpressInstall the CORS Middleware ModuleCreate the Entire Server File with ExpressLaunch Your Server and TestStream to the UI with FetchDecode the Response Chunks and RenderSummaryExercises
3. PYTHON ESSENTIALS FOR LLMS AND APIS
Installing Python and LibrariesCalling an LLM via an APICreate the FastAPI ServerUnderstand the ComponentsSet Up the FastAPI ServerSend a Prompt to the Local LLMLaunch Your Server and TestSummaryExercises
PART II: PROMPT ENGINEERING
4. FUNDAMENTALS OF PROMPT ENGINEERING
Programming vs. PromptingA Clear Use Case for an LLMEvaluating Your Use CasePrompting BasicsPrompt ElementsContext WindowsTokenizationAvoiding Runaway CostsAccuracy and PerformanceChoosing a ModelSummaryExercises
5. PROMPT ENGINEERING TECHNIQUES
Instruction PromptingData ExtractionPersonally Identifiable Information RemovalSynthetic Data GenerationPersona PromptingZero-, One-, and Few-Shot PromptsChain-of-Thought PromptingPrompt ChainingPrompt Tips and Best PracticesFocus on the Positive, Avoid the NegativeCollaborate with the Model on Your PromptTreat Your Prompts Like CodeExperiment with Prompt VariationsSummaryExercises
6. PROMPT ENGINEERING IN CODE
Choosing a LibraryConnecting to a Hosted APILLM Output ConfigurationOutput LengthSampling ControlsPrompt TemplatesCreating a Template with JinjaSeparating Prompts and Code with JinjaUsing Dynamic Prompt Templates with JinjaMessages and RolesUnderstanding Chat APIsIncluding System MessagesCreating a Type for Messages and RolesConversation HistoryMaintaining Conversation History on the ClientTrimming the Conversation HistorySummarizing the Conversation HistoryChoosing Conversation History OptionsStructured OutputDefining a JSON Schema with PydanticUsing InstructorExtracting Data with InstructorPrompting in ProductionRequest RetriesGuardrails and InjectionToken Counts and Usage CostsSummaryExercises
PART III: VECTOR DATABASES AND RAG
7. VECTOR DATABASES IN PRACTICE
Helping Machines Understand Our WorldPowering a Better Search ExperienceUnderstanding Vector EmbeddingsChoosing an Embedding ModelDissecting Model NamesHow Vector Database Retrieval WorksUsing a Vector Database for Meaningful Search ResultsCreating the Product CollectionInserting Products into the Vector DatabaseQuerying the Vector DatabaseUsing a Vector Database to Make RecommendationsDefining the Data ClassesRetrieving the RecommendationsIdentifying the Most Relevant ResultsFiltering on MetadataSummaryExercises
8. DESIGNING A RETRIEVAL-AUGMENTED GENERATION SYSTEM
What Is Retrieval-Augmented Generation?Connecting the Knowledge BaseUsing Context Retrieval with a Vector DatabaseUnderstanding the RAG PipelineSetting Up the EnvironmentChunking and Loading a Vector DatabaseUnderstanding Chunking BasicsChoosing a Chunking StrategyChunking the DocumentationLoading Chunks into the Vector DatabaseCompleting the RAG Pipeline with an LLMCreating the Prompt TemplatesPrompting an LLM with ContextCreating the Service APIImproving Trust with CitationsAdvanced RAG TopicsAlternative Indexing and Retrieval StrategiesArchitectural VariantsProduction-Level RefinementsData Quality and SecuritySummaryExercises
PART IV: ADAPTING MODELS TO REAL-WORLD TASKS
9. WHY AND WHEN TO CUSTOMIZE A MODEL
Building a New Model vs. Fine-TuningFine-Tuning StrategiesSelf-Supervised Fine-TuningSupervised Fine-TuningReinforcement Learning Fine-TuningFine-Tuning MethodsNo-Code Fine-TuningTechnical Fine-TuningThe Fine-Tuning ProcessSummary
10. PREPARING DATA FOR FINE-TUNING
Gathering the Raw DataStructuring the Raw DataEncoding the LabelsSplitting the DataCreating a Labeled DatasetCreating a DatasetCreating the Train, Validate, and Test DatasetsCombining the DatasetsVerifying the DatasetFixing the LabelSummaryExercises
11. FINE-TUNING MODELS IN PRACTICE
Choosing Models for Fine-TuningEstablishing a BaselineFine-Tuning a Classification ModelEvaluating the Fine-Tuned Classification ModelFine-Tuning an LLMWhy Not Prompt Engineering?Introduction to PEFT and LoRARegular Fine-Tuning vs. PEFTLoRAPrompting an LLMUnderstanding Instruction TemplatesCustomizing Generated Response FormatControlling Where Generation StartsBuilding the DatasetTraining the LLM Using LoRAReducing ParametersConfiguring LoRATrainingEvaluating the Fine-Tuned ModelSummaryExercises
PART V: BUILDING AGENTIC SYSTEMS
12. FROM WORKFLOWS TO AUTONOMOUS AGENTS
What Are AI Agents?Why Do We Need Agents?Traditional WorkflowsAI WorkflowAgentic WorkflowSummary
13. BUILDING AN AUTONOMOUS AGENT
Setting Up Your Agent Development EnvironmentBuilding Your First AgentAgents and Prompt EngineeringSummaryExercises
14. EXTENDING AGENTS WITH TOOLS
Building a ToolWriting Effective ToolsBuilding Reusable ToolsModel Context ProtocolMCP ConceptsMCP ComponentsMCP Transport ProtocolsBuilding an MCP ServerYour First MCP ServerBuilding a Special CalculatorDefining Custom ToolsRunning the MCP ServerTesting with PostmanIntegrating with Claude DesktopMCP TransportsBuilding Tools with Complex SchemasSummaryExercises
AFTERWORD
INDEX

Content preview from The Developer's Guide to AI

8DESIGNING A RETRIEVAL-AUGMENTED GENERATION SYSTEM

LLMs arrive with impressive capabilities. But they’re also blank slates when it comes to your own data. They know only what was in their training set and have no built-in awareness of your internal wikis, support tickets, email, or database records. Yet, as our startup’s AI support agent has shown, connecting an LLM to your private content unlocks powerful new possibilities for automation and efficiency.

Retrieval-augmented generation (RAG) is the architectural pattern that makes this practical. All its variations have the same intent: delivering exactly the right pieces of your private data ...

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.

Read now

Unlock full access

More than 5,000 organizations count on O’Reilly

O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.

Julian F.

Head of Cybersecurity

I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.

Addison B.

Field Engineer

I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.

Amir M.

Data Platform Tech Lead

I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.

Mark W.

Embedded Software Engineer

Publisher Resources

ISBN: 0642572230333Errata Page

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills

The Developer's Guide to AI

by Jacob Orshalick, Jerry M. Reghunadh, Danny Thompson

8DESIGNING A RETRIEVAL-AUGMENTED GENERATION SYSTEM

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.