Mishra A. - Mastering Multi-agent Systems with Azure OpenAI [2026, PDF/EPUB, ENG]

Книги и журналы » Компьютерная литература » Программирование (книги)

Mastering Multi-agent Systems with Azure OpenAI

Год издания: 2026
Автор: Mishra A.

Издательство: Apress
ISBN: 979-8-8688-2854-6
Язык: Английский

Формат: PDF/EPUB
Качество: Издательский макет или текст (eBook)
Интерактивное оглавление: Да
Количество страниц: 467

Описание: Contains Real-world use cases and implementation for Smart Travel Agent and Smart Restaurant
Code illustration in both Python and C#
Discusses implementation of multi-agent interoperability with Model Context Protocol (MCP) and A2A
This book is your definitive guide to building intelligent, autonomous, and collaborative AI systems on the Microsoft ecosystem. It covers the rapid evolution of Agentic AI and multi-agent frameworks, empowering developers, architects, and AI practitioners to design and implement real-world, production-ready multi-agent solutions using Azure OpenAI and emerging protocols such as Model Context Protocol (MCP) and A2A.
Starting with the fundamentals of Generative AI, Large Language Models (LLMs), and Agentic AI, you’ll dive into the practical aspects of building agents using Microsoft’s Agent Framework. You’ll learn how to architect, fine-tune, and secure Azure OpenAI services, build structured multi-turn conversations, and develop specialized agents that can communicate, collaborate, and adapt autonomously. The book goes beyond theory by walking you through hands-on, industry-grade projects—such as Smart Travel Agent and Smart Restaurant—to demonstrate how multi-agent systems can transform real-world scenarios. Moving ahead, you will also explore advanced topics like Retrieval-Augmented Generation (RAG), implementing Model Context Protocol for interoperability, and ensuring ethical, responsible AI design.
Large Language Models are pre-trained with a huge amount of data. They do not learn from prompts. The agentic solutions in an enterprise work in the domain context of the organization. For example, an agent in a restaurant chain should respond with food menu options and prices designed for its restaurants. The Large Language Model does not know the restaurant menu and prices, as it is not trained with them. So the agents in the enterprise should work on a restaurant-specific knowledge base and not the pre-trained knowledge base of the Large Language Model. Retrieval-Augmented Generation (RAG) is the solution for this.
Retrieval-Augmented Generation (RAG) helps agentic solutions to act and respond based on knowledge from external databases instead of relying on the Large Language Model’s pre-trained data. The external databases on which the RAG-based agents work are called vector databases. The vector databases store the knowledge as numerical representations called embedding instead of text. The agents perform a semantic search where the queries are also converted into embeddings and are matched against the embeddings stored in the vector database. This mechanism facilitates search based on intent and meaning instead of plain text search, as in traditional search queries that run on traditional databases.
By the end of this book, you’ll not only understand the technical depth of multi-agent architectures but also gain the practical expertise to design, deploy, and scale intelligent, context-aware AI ecosystems on Azure.
What You Will Learn:
- Multiple autonomous agents based Agentic AI solutions using Azure AI and Microsoft Agent Framework
- Multi-Agent collaboration using Model Context Protocol (MCP) and A2A
- RAG-Enabled Agents using Azure AI Search
- Best practices and pitfalls to build Agentic AI solutions along with social and ethical considerations.
[spoiler="Примеры страниц (скриншоты)"]

[/spoiler]
[spoiler="Оглавление"]
About the Author xi
About the Technical Reviewer xiii
Acknowledgments xv
Introduction xvii
Chapter 1: Introduction 1
Chapter 2: Azure OpenAI 23
Chapter 3: Microsoft Foundry 59
Chapter 4: Building AI Agents 117
Chapter 5: Building Multi-agent Systems Using Model Context Protocol and A2A 203
Chapter 6: Building Retrieval-Augmented Generation (RAG)–Based Agent Systems 261
Chapter 7: Requirements and Design for Smart Travel Agent 311
Chapter 8: Smart Travel Agent Using LangGraph and LangChain 321
Chapter 9: Requirements and Design Smart Restaurant 379
Chapter 10: Smart Restaurant Using Microsoft Agent Framework and Azure AI Search 387
Chapter 11: Agentic AI Design and Implementation Guidance 447
Index 455
[/spoiler]