Agentic AI Second Edition

Agentic AI is a practical and comprehensive guide to understanding and building intelligent AI agents capable of reasoning, planning, using tools, interacting with external systems, and completing complex tasks.

The book begins with the fundamentals of Artificial Intelligence, Generative AI, and Large Language Models (LLMs) and gradually progresses to advanced Agentic AI concepts. Readers will learn how AI agents use models, tools, memory, knowledge, workflows, and decision-making processes to perform tasks with greater autonomy.

The book provides practical coverage of LangChain and LangGraph, including state management, nodes, edges, conditional routing, ToolNode, tool calling, human-in-the-loop workflows, middleware, and multi-agent systems.

With practical examples, real-world use cases, and hands-on projects, this book helps readers develop modern AI-powered applications and understand the architecture behind intelligent and autonomous AI systems.

What You Will Learn
Introduction to Artificial Intelligence
Introduction to Generative AI
Generative AI vs Agentic AI
Large Language Models (LLMs)
How LLMs work
Tokens and context windows
Model parameters and temperature
Prompt Engineering
Prompt design and optimization
Zero-shot, one-shot, and few-shot prompting
Structured output generation
Introduction to AI Agents
What is Agentic AI?
Agentic AI architecture
Components of an AI Agent
LLMs as reasoning engines
Planning and decision-making
Reasoning and task decomposition
Agent memory
Short-term and long-term memory
Tool usage and external systems
Tool Calling
Function Calling
Custom tools
API and database integration
Web search and external information retrieval
Introduction to RAG
Retrieval-Augmented Generation architecture
Document processing
Text splitting and chunking
Embeddings and vector databases
Semantic search
Retrievers
Building knowledge-based AI agents
Introduction to LangChain
LangChain architecture and components
Prompt templates
Chains and Runnables
LangChain agents
Tools and tool integration
Agent memory
Introduction to LangGraph
LangGraph architecture
Graph-based AI workflows
Nodes and edges
State management
Conditional routing
Loops and iterative workflows
ToolNode
Human-in-the-loop workflows
Middleware
Agent control and customization
Multi-step AI workflows
Multi-Agent Systems
Agent collaboration
Agent-to-agent communication
Supervisor and worker agent patterns
Autonomous AI workflows
AI Agent planning and execution
Error handling and recovery
Agent evaluation and monitoring
LangSmith and AI application observability
Hugging Face and LLM integration
Working with modern AI APIs
Database and API integration
AI-powered chatbots
Research agents
Customer support agents
Data analysis agents
Coding agents
Document analysis agents
Automation agents
Security and responsible Agentic AI
Prompt injection and tool security
Controlling agent permissions
Hallucination reduction
Cost and performance optimization
Deploying Agentic AI applications
Real-world Agentic AI projects