book

Generative AI-Driven Application Development with Java: Leveraging Large Language Models in Modern Java Applications

Name: Generative AI-Driven Application Development with Java: Leveraging Large Language Models in Modern Java Applications
Author: Satej Kumar Sahu
ISBN: 9798868816093

by Satej Kumar Sahu

January 2026

Intermediate to advanced

713 pages

13h 39m

English

Apress

Read now

Unlock full access

Generative AI-Driven Application Development with Java
Introduction
Who Should Read This Book
Note to Readers
Acknowledgments
Table of Contents
About the Author
About the Technical Reviewer
1. Megabrains 101: Generative AI and LLMs Unboxed
Understanding Gen AIUnderstanding LLMsUnderstanding the Role of Hugging Face in the LLM WorldEcosystem Around LLMsChallenges and Considerations in Using LLMsSetting the Stage: LLMs in Java DevelopmentConclusion
2. First Contact: “Hello, LLM” with Spring Boot
IntroductionWhat Is OpenAI?Creating Your First Spring Boot LLM ApplicationExploring the OpenAI Chat APIIntegrating OpenAI API with Spring BootConclusion
3. The Transformer Saga—From Attention to Fine-Tuning
IntroductionImportance of LLMsHow Large Language Models WorkTransformerEvolution of TransformersFine-Tuning LLMsSupervised Fine-TuningDialogue and Instruction TuningSafety TuningReinforcement Learning from Human FeedbackParameter-Efficient Fine-TuningInference Optimization TechniquesQuantizationDistillationOutput-Preserving OptimizationsSpeculative DecodingBatching and ParallelizationConclusion

4. Bring Your Own Model: Self-Hosting with Ollama
IntroductionSelf-Hosting LLMsOllamaInstalling OllamaRunning a Locally Hosted LLMExploring the APIs Exposed by OllamaIntegrating Ollama with Your Existing ApplicationConclusion
5. Power Tools: LangChain4j Quick-Start
IntroductionWhat Is LangChain4j?Installation and Setup of LangChain4jInitial Project Setup: Creating Your First LangChain4j ApplicationConclusion
6. Integrating LLMs with Java Applications
IntroductionPrompt EngineeringContext into Conversation MemorySwitching Between LLM ProvidersHandling LLM OutputsSecurity ConsiderationsConclusion
7. From Chatty to Clever: Retrieval-Augmented Generation
IntroductionContextuality in the LLM EcosystemRAG ParadigmEmbeddingsVector StoreRAG in ActionConclusion
8. Spring AI Ninja Moves
IntroductionWhat Is Spring AI?Features of Spring AIBoot Up an Application Using Spring AIUsing OpenAI API Through Ollama API in Spring AIBuilding a RAG-Based ApplicationRetrievalConclusion
9. Prompt Alchemy: Patterns That Make Models Look Smarter
IntroductionWhat Is a Prompt?What Is Prompt Engineering?Prompt Engineering PatternsCreating Effective PromptsKey Components of a PromptExamples of Expected Output FormatSimple Prompting TechniquesAdvanced Prompting TechniquesMicrosoft’s Framework for Prompt Creation and OptimizationBest Practices for Prompt EngineeringUsing Prompt Engineering in Spring AITemperatureOutput Length (MaxTokens)Structured Response FormatModel-Specific OptionsPrompt Engineering TechniquesZero-Shot PromptingOne-Shot and Few-Shot PromptingSystem, Contextual, and Role PromptingSystem PromptingRole PromptingContextual PromptingStep-Back PromptingChain of ThoughtSelf-ConsistencyTree of ThoughtsAutomatic Prompt EngineeringCode PromptingNotesConclusion
10. Swiss-Army LLMs: Tool Calls in Spring AI
IntroductionWhat Are Tools?Understanding the Flow of Control with ToolsUnderstanding Tool Calling with Spring AIStandard Practices in Tool CallingStarting with a Simple Use of ToolsDeclarative Specification: @ToolExploring an Advanced Use Case of ToolsConclusion
11. Agents Assemble! Building Autonomous Workflows
IntroductionAgentic AIKey Characteristics of an AgentTypes of AgentsApplications of AgentsDefining ConceptsComponents of Agentic SystemsModelsToolsKnowledge and MemoryGuardrailsOrchestrationAgentic Workflows and PatternsChain WorkflowWhen to Use Chain WorkflowWhat’s Happening in the Workflow?Parallelization WorkflowWhat’s Happening in the Process?Routing WorkflowWhen to UseWhat’s Happening in the Process?Key Insights:Agentic WorkersOrchestrator-WorkersWhen to UseOrchestratorWorkersWorkflow ClassWhat’s Happening in the Process?Key InsightsEvaluator-OptimizerWhen to UseEvaluatorOptimizerWorkflow ClassWhat’s Happening in the Process?Key InsightsAgent Lifecycle in Spring AISpring AI’s Implementation AdvantagesModel PortabilityStructured OutputConsistent APIBest Practices and RecommendationsStart SimpleDesign for ReliabilityConsider Trade-OffsAdvisors API in Spring AIConclusion
12. Quarkus + LangChain4j: Lightning-Fast Gen AI
IntroductionIntegration of LangChain4j in QuarkusPrompt Engineering in QuarkusCrafting Prompts with AnnotationsExposing an AI-Powered REST EndpointUsing Tools in QuarkusUsing ChatMemory in QuarkusDefining a ChatMemoryProviderEnabling Memory in Your AI ServiceUsing the RAG Pattern in QuarkusIngest Documents into Your Redis Vector StoreDefine Your RetrievalAugmentorEnable RAG in Your AI ServiceConclusion
13. Jlama and Friends: Hosting Models the Java Way
IntroductionWhat Is Jlama?Jlama: Quick StartUse Jlama to Host an LLMQuantize a Model Using JlamaUse Jlama with Quarkus to Build a Gen AI ApplicationConclusion
14. Seeing Is Believing: Multimodal LLMs and Image Hacking
IntroductionTooling in Java for Image ProcessingMultimodal LLMsBuilding an Application to Process Flight TicketsGenerating Images from User RequestOther multimodal formatsConclusion
15. Does It Even Work? Testing and Evaluating LLM Apps
IntroductionEvaluation of LLM ApplicationsTools Available for Evaluation and TestingTesting and Evaluating Using Spring AIConclusion
16. Cloud Power-Ups: Bedrock, Vertex, and Azure OpenAI
IntroductionState of Gen AI in the CloudGenerative AI in AWS: Amazon BedrockAccessing a Foundational Model in Bedrock PlaygroundBuilding a Sample Project Using Amazon Bedrock SDKUsing AWS SDK to Connect with Amazon BedrockExtending and Integrating Your Application with GuardrailsAmazon Bedrock Guardrails: Enabling Responsible AI at ScaleExtending and Integrating RAG into Your ApplicationReengineering Your Application Using Spring AI Integration with Amazon BedrockGenerative AI in GCPGenerative AI in AzureConclusion
17. Talking in Protocols: The MCP Revolution
IntroductionProtocols in the LLM EcosystemModel Context ProtocolAgent Communication ProtocolAgent-to-Agent ProtocolAgent Network ProtocolWhat Is MCP Protocol?Role of MCPMCP Using Spring AIMCP ClientMCP ServerSpring AI MCP IntegrationCapabilities of the MCP Server in Spring AITool Exposure and ManagementResource ManagementPrompt TemplatesCompletionsRoot Change MonitoringCreating Spring AI-Based MCP ServerCreating a Spring AI-Based MCP ClientIntegrating MCP with QuarkusConclusion
18. Can You See Me Now? Observability for LLM Pipelines
IntroductionWhy Observability Matters in LLM EvaluationCore Observability SignalsBest PracticesIntegrating Observability in LLMs Through Spring AIObserving ChatClient RequestsObserving Chat Model RequestsConclusion
19. Native-Speed Machine Learning in Java: DJL, ONNX, and JNI
IntroductionUsing Hugging Face APIs in JavaAccessing Hugging Face Models via an APICreating Java Wrappers for InferenceIntegration via FrameworksSpring AIQuarkus + LangChain4jDeep Java Library (DJL)Hosting and Inference Using DJLUnderstanding the Linnerud DatasetDataset StructurePhysiological Data: Input Features (X)Exercise Data: Target Variables (Y)Use in Machine LearningWhy Use This Toy Dataset?Building an ONNX Regression Model with the Linnerud DatasetStep-by-Step: Training and Exporting the ModelStep 1: Load and Prepare the DatasetStep 2: Train a Regression ModelStep 3: Convert the Model to ONNX FormatStep 4: Test Inference in PythonWhy This Workflow WorksIntegrating LLMs Using Java Native InterfaceWhy Use JNI for LLM Integration?Sample JNI Integration with a Native LLMBehind the ScenesWhen to Use JNI for LLMsConsiderationsRunning Inference in Java with ONNX RuntimeWhy Use ONNX Runtime in Java?Getting Started with ONNX Runtime for JavaKey ConceptsUse CasesModel Training and Fine-Tuning Using DL4JKey Components of the DL4J SuiteTraining and Fine-Tuning ModelsWhy Use DL4J?Open Source and CommunityNatural Language Processing Using Apache OpenNLPGetting StartedWhy Use OpenNLP?Conclusion
20. Architectures of Tomorrow: From Monoliths to Modular Minds
IntroductionArchitectural and Scaling Trends in Generative AI: From Monoliths to Modular IntelligenceSmarter Scaling: From Large Models to Efficient ArchitecturesRAG: Memory-Enhanced Systems for Real-Time ReasoningAgentic AI: Modular Orchestration Through Intelligent AgentsProtocol-Driven Interoperability: MCP, ACP, A2A, and ANPPerformance-Aware Design: Beyond the ModelA New Architecture for Scalable IntelligenceEcosystem Showdown: Open-Source vs. Cloud-Hosted ModelsNavigating the Challenges of Generative AI: What Developers Must Know and DoPrompt Injection, Jailbreaks, and Model ExploitsEthical Use and Bias MitigationData Privacy and LeakageGovernance and ComplianceEnergy Use and Environmental CostModel Hallucinations and Factual IntegrityStaying Current: The Developer’s PlaybookThe Evolving Pathway to LLM Adoption: From Prompts to Agentic AIPhase 1: Prompt Engineering and Model TrialsPhase 2: Fine-Tuning and Domain AdaptationPhase 3: Model Training from Scratch (Advanced Tier)The Rise of Agentic AIManaging the LLM Development LifecycleLLMOps: Operationalizing the Lifecycle of Generative AI SystemsCore Components of LLMOpsLLMOps in Practice: From Playground to ProductionThe Road AheadBridging the Gap: Migrating from DevOps to LLMOps ThinkingFrom Deterministic Code to Probabilistic OutputsFrom CI/CD Pipelines to Continuous EvaluationFrom Logs and Metrics to Semantic ObservabilityFrom Source Code to Prompt + Model as CodeFrom Feature Testing to Use-Case AuditingFrom Code Deployments to Model Lifecycle ManagementFrom User Feedback to Human-in-the-Loop LearningFinal Thought: It’s Still About CultureThe Evolving Culture and Practices in Java’s Machine Learning EcosystemJava Meets Machine Learning: A Culture Shift in the MakingDJL: Lowering the Barrier for Java DevelopersEclipse Deeplearning4j: Full-Stack Deep Learning on the JVMJlama and the Expanding ToolchainWhere the Culture Is HeadingA Rising Force in AI DevelopmentEmerging Frontiers: Multimodality, Agents, and Domain SpecializationRegulations and ActsWhat This Means for BuildersGovernance, Ethics, and Future OutlookThere’s Just Too Much Happening. What Should I Do?Focus on Fundamentals, Not the FOMOYou Already Have Transferable KnowledgeEmbrace the Culture of LearningAdaptation Is the Skill of the DecadeConclusion
Index

Content preview from Generative AI-Driven Application Development with Java: Leveraging Large Language Models in Modern Java Applications

S. K. SahuGenerative AI-Driven Application Development with Javahttps://doi.org/10.1007/979-8-8688-1609-3_9

9. Prompt Alchemy: Patterns That Make Models Look Smarter

Satej Kumar Sahu¹

(1)

Berhampur, Odisha, India

The goal of this chapter is to explore the art of prompt engineering using Spring AI.

Introduction

This chapter into the fascinating world of prompt engineering using Spring AI. As artificial intelligence continues to evolve, the ability to effectively communicate with large language models (LLMs) has become a crucial skill. A prompt, the input text given to an AI model, plays a significant role in shaping the output. By carefully crafting these prompts, ...

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: 9798868816093Purchase Link Publisher Website

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills

Generative AI-Driven Application Development with Java: Leveraging Large Language Models in Modern Java Applications

by Satej Kumar Sahu

9. Prompt Alchemy: Patterns That Make Models Look Smarter

Introduction

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.