Practical AI Agent Development with ADK: From Concept to Deployment

Hwang, SamChung 지음 | Hwang, JunSik 그림

넷스쿨 2025년 05월 24일 출간

작품소개

책소개

Practical AI Agent Development with ADK: From Concept to Deployment is a hands-on guide to building autonomous AI agent systems with Google’s Agent Development Kit (ADK). Written for AI professionals and software developers, it provides a roadmap for designing intelligent agents and multi-agent workflows that perceive, decide, and act autonomously. It focuses on real-world use cases and introduces Google’s ADK as an innovative framework to streamline agent development from concept to deployment.
Structure and Learning Flow
The book is structured in four parts:
Part I – Foundations: Explains what AI agents are, why multi-agent systems matter, and introduces ADK’s core architecture.
Part II – Building Agents and Workflows: Shows how to create your first AI agent and then compose agents into workflows. You learn to manage state and handle errors for reliability.
Part III – Advanced Collaboration: Explores orchestrating multiple specialized agents (e.g. decision-tree strategies for complex tasks) and integrating external data sources, APIs, and tools to extend capabilities.
Part IV – Projects and Deployment: Presents full case-study projects that apply everything learned. You will build end-to-end solutions (such as a document Q&A assistant and an e-commerce automation pipeline) and learn how to containerize and deploy your agent system to the cloud with CI/CD, monitoring, and security measures.
Key Concepts and Tools
AI Agents & Multi-Agent Workflows: Understanding autonomous AI agents and how multiple agents coordinate in workflows to tackle complex tasks more effectively than a single agent.
Modular Architecture (Google ADK): Using ADK’s modular design to build agents as reusable components and orchestrate their interactions. ADK is model-agnostic, so you can plug in different language models (e.g. Google’s Gemini) as needed.
Integration with APIs and Tools: Connecting agents to external systems and data sources. Learn how to equip agents with access to databases, web services, and custom APIs (for tasks like searching documents or checking inventories), and how to use advanced language models as reasoning engines.
Cloud Deployment: Deploying AI agents in production environments. You’ll learn how to containerize an agent application and deploy it on cloud platforms like Google Cloud Run, set up continuous integration/delivery, and add monitoring and security for a production-ready solution.
Hands-On Practice and Projects
Throughout the book, each chapter includes code examples and encourages you to follow along with a provided code repository. Early exercises let you build confidence with ADK basics, and later chapters tackle more complex multi-agent scenarios. Finally, the book provides immersive projects. You will implement a document question-answering agent that can search a private knowledge base for answers, and a multi-agent e-commerce workflow that automates steps from product inquiry to order fulfillment. Each project comes with complete code and deployment guidance, demonstrating how to move from a working prototype to a scalable, cloud-deployed service.
Audience and Benefits
This guide is ideal for AI engineers, ML practitioners, and software developers who want to create the next generation of AI-driven applications. If you’re looking to go beyond basic chatbots and harness multiple AI agents working in concert, this book will show you how. It doesn’t stop at prototyping – you’ll learn to build robust agents and deploy them confidently to production. By the end of Practical AI Agent Development with ADK, you will have the knowledge and hands-on experience to design, build, and deploy intelligent agents using Google’s ADK. You’ll be equipped to create robust multi-agent systems for real-world use and take your AI agent solutions from concept to deployment.

목차

Practical AI Agent Development with ADK: From Concept to Deployment
Prologue
Part I: Foundations of AI Agents and the Google ADK Architecture
Chapter 1: Understanding AI Agents and Their Role in Automation
Chapter 2: Language Models, Development Frameworks, and Setting Up Your ADK Environment
Chapter 3: Google ADK Architecture and Core Components

Part II: Building Multi-Agent Workflows with Google ADK
Chapter 4: Building Your First AI Agent with Google ADK
Chapter 5: Designing Workflows and Managing State in ADK
Chapter 6: Handling Errors and Ensuring Robustness in AI Agent Workflows
Part III: Collaboration Strategies and Tool Integration
Chapter 7: Improving Decision Accuracy with a Decision-Tree Approach
Chapter 8: Connecting AI Agents to External Systems and Data Sources

Part IV: Hands-On Projects for Workflow Automation
Chapter 9: Project – Integrating AI Agents with Vertex AI and Cloud APIs
Chapter 10: Project – Automating E-commerce Workflows with AI Agents
Chapter11: Project: Self-RAG implementation

Part V: Optimizing and Deploying AI Agents in Production
Chapter 12: Optimizing and Scaling AI Agent Systems
Chapter 13: Deploying and Orchestrating AI Agents on Google Cloud(Windows & ADK Edition)
Conclusion
판권

책 속으로

Overview
Practical AI Agent Development with ADK is a comprehensive guide to building intelligent agents and multi-agent systems using Google’s Agent Development Kit (ADK). Google’s ADK is a new open-source framework that simplifies the development of both single-agent and multi-agent applications. The book emphasizes a hands-on approach: from fundamental concepts of AI agents to deploying production-ready agent workflows on Google Cloud. Readers progress through every step of the development lifecycle – from initial concept and prototyping to scaling up and cloud deployment – with plenty of code examples and real-world use cases.

Book Structure and Key Topics
Part I – Foundations: Covers what AI agents are, how they work (sensing, reasoning, acting), and why multi-agent systems are needed for complex tasks. It establishes the conceptual foundation and sets up the development environment.
Part II – Building Agents with ADK: Guides you through creating your first AI agent using ADK, then expanding it into multi-step workflows. You learn how to design agent logic, manage state, and handle errors for reliability.
Part III – Collaboration & Integration: Explores advanced multi-agent strategies. You will orchestrate multiple specialized agents (e.g. a decision-tree approach for better decisions) and integrate them with external resources like APIs, databases, and live data to build richer systems that function in real-world environments.
Part IV – Hands-On Projects: Shifts from theory to practice with end-to-end projects applying ADK. You learn to connect agents to enterprise cloud services, enable them to process documents intelligently, and automate multi-step business workflows. These projects span enterprise integration, document analysis, and business process automation, reinforcing earlier concepts in a tangible way.
Part V – Production Deployment: Focuses on optimizing and deploying AI agents. You learn performance tuning (reducing latency, scaling concurrency) and how to deploy the multi-agent system on Google Cloud with CI/CD pipelines, monitoring, and security best practices.

Distinctive Features
The book’s practical, code-first approach sets it apart. Every chapter is example-driven, with code and a companion GitHub repository for hands-on practice. This ensures you gain practical experience, not just theoretical knowledge. Another highlight is the focus on Google’s ADK itself. The book emphasizes ADK’s innovative design as a “powerful, modular framework that puts production-ready multi-agent systems within reach of every developer”. It shows how ADK’s components can be extended: you can compose sub-agents into larger workflows, plug in any suitable language model (Google Gemini or others), and integrate new tools or APIs with minimal effort. Combined with guidance on proper deployment and scaling, these features ensure the techniques you learn are readily applicable to real-world projects.

Who Will Benefit
This book is ideal for AI/ML engineers and software developers who want to transition from basic prototypes to production-ready AI agent systems. Assuming a background in Python and elementary AI knowledge, it provides step-by-step guidance to implement intelligent assistants, chatbots, and workflow automation agents that operate reliably in real-world conditions. By the end of the journey, you will be “fully equipped to harness the power of AI agents in your own projects”

– ready to leverage Google’s ADK to turn AI concepts into deployed, scalable solutions.

출판사 서평

If you’ve ever wanted to build your own intelligent AI agents but weren’t sure where to start, Practical AI Agent Development with ADK: From Concept to Deployment is the book you’ve been waiting for. This friendly, comprehensive guide by SamChung Hwang takes readers on an exciting journey into the world of AI agents. The writing is enthusiastic and approachable, making advanced concepts feel easy. The book brilliantly balances practical focus with accessible explanations, all while showcasing the powerful capabilities of Google’s Agent Development Kit (ADK). It’s a must-read for anyone eager to understand or build intelligent AI agents.

A Practical and Accessible AI Guide for Everyone
One of the greatest strengths of this book is its accessibility. SamChung Hwang writes in a clear, conversational style that makes complex AI concepts easy to grasp for general readers and developers alike. The book starts from the ground up, explaining what AI agents are and why they matter before gradually introducing more advanced topics. Throughout each chapter, the focus remains on real-world understanding – every concept comes with code examples or analogies that demystify the tech jargon. The result is a practical guide where you learn by doing. Each chapter is code-centric and example-driven, so you’re never just reading theory – you’re building something tangible.

Harnessing the Power of Google’s ADK
At the heart of this guide is Google’s powerful Agent Development Kit. ADK is a revolutionary open-source framework that simplifies the creation of both single-agent and multi-agent systems. Practical AI Agent Development with ADK demystifies ADK’s cutting-edge features and shows you how to leverage them step by step. You’ll discover how ADK’s modular architecture and workflow orchestration let you connect multiple specialized agents into a well-coordinated team. Each agent can handle different subtasks seamlessly, sharing information to achieve complex goals together. The book highlights how ADK integrates with external tools and data sources (APIs, databases, web services) and remains model-agnostic – meaning you can plug in any AI language model of your choice. By following the examples, you’ll quickly grasp how ADK empowers developers to build intelligent agents that perceive information, make decisions, and take actions autonomously. The author’s enthusiasm for ADK shines through, inspiring you to experiment with this powerful toolkit in your own projects.

Hands-On Projects from Concept to Deployment
What truly sets this book apart is its abundance of real-world examples and step-by-step projects. Practical AI Agent Development with ADK doesn’t stop at teaching theory or presenting isolated code snippets – it walks you through complete, end-to-end AI agent solutions. From designing an AI agent’s basic reasoning abilities to orchestrating a team of agents working together, you’ll follow detailed case studies that bring the concepts to life. For instance, you’ll build an automated e-commerce order processing pipeline and a document Q&A assistant, each with clear instructions and sample code you can follow or adapt. Every project is broken down into manageable steps, allowing you to code along and modify things as you learn. The guide even covers how to take your AI agents from concept all the way to a deployed application in a production environment. You’ll learn how to containerize your agent applications and deploy them on Google Cloud Run, complete with best practices for continuous integration (CI/CD), monitoring, and security. By the end of these hands-on chapters, you won’t just understand intelligent agents in theory – you’ll have built and deployed actual AI applications yourself.

What You Will Learn
By reading this book, you will gain the skills to:
- Design and implement intelligent agents that can reason, use tools, and maintain context autonomously.
- Compose multi-agent workflows where multiple AI agents collaborate to tackle complex tasks together.
- Integrate external systems and data – connecting your agents to APIs, databases, and cloud services to expand their capabilities.
- Build real-world AI applications (like chatbots, automated workflows, or Q&A assistants) with complete, customizable sample code provided.
- Deploy and manage AI agents in production, using cloud platforms (such as Google Cloud Run) with CI/CD pipelines, monitoring, and security best practices.

A Must-Read for Aspiring AI Innovators
In conclusion, Practical AI Agent Development with ADK: From Concept to Deployment is an invaluable resource for anyone who wants to dive into the world of intelligent AI agents. Whether you’re a seasoned software engineer or a curious newcomer, the book’s step-by-step, enthusiastic approach will leave you empowered and inspired. SamChung Hwang has crafted a guide that not only teaches you about AI agents but also gives you the confidence to build your own. This book truly deserves a spot on the shelf of anyone looking to create the next generation of AI-driven applications. By the time you turn the last page, you’ll be fully equipped to harness the power of AI agents in your own projects – and have a lot of fun along the way!

작가정보

저자(글) Hwang, SamChung

SamChung Hwang holds a B.S. in Computer Engineering from Seoul National University (February 1988) and earned his Ph.D. in Computer Science with a specialization in Artificial Intelligence from KAIST (August 1994). He began his software-engineering career in 1994 at HandySoft, where he led the development of intranet groupware—designing both the approval workflow and core engine.

Over the next decade, Dr. Hwang built his technical leadership at CyberBank and ThinMultimedia, serving as Development Team Lead and Research Director for terminal-side applications such as SW-based VOD and SMIL players. He then joined Samsung Electronics’ Wireless Division as a Senior Researcher and Development Manager, driving smartphone software projects. At Naver, he oversaw the Development Lab for Service Management Systems and Map Services, guiding large-scale, mission-critical platforms.

Today, as CEO of Netschool, Dr. Hwang combines his decades of software development and AI expertise with real-world investment insight—researching and building AI-driven investment models across equities and real estate. In parallel, he delivers investment courses and consulting services, and continues to explore innovative business applications of AI agents through both research and teaching.

그림/만화 Hwang, JunSik