Build AI Agents using MCP

Overview

Get hands-on designing secure, intelligent AI agent workflows using the Model Context Protocol (MCP) in this labs-driven course. You’ll see how AI systems connect to external tools, services, and data sources. You’ll learn how those connections can be designed to stay safe and predictable using structured permissions, user prompts, and validation workflows. And in hands-on labs, you’ll build agents that reason, retrieve information, and carry out tasks while maintaining security and control. You’ll also work with permission enforcement models, JSON-schema-based elicitation, auditing concepts, and real-world security scenarios. You’ll explore how MCP works and why secure design decisions matter in practice. Plus, you’ll break down user requests, shape safe execution flows, and reduce the risk of unintended actions. Finally, you’ll plan and test a complete MCP-driven agent workflow, showing how usability, capability, and security come together in a real implementation. This course is designed for professionals in development, architecture, automation, or AI-powered applications who want hands-on, practical experience building responsible AI workflows.

Syllabus

Getting Started with MCP

In this module, you will gain a hands-on introduction to the Model Context Protocol (MCP). You will explore what MCP is, why it is used, and how it solves challenges compared to traditional APIs and tool-calling approaches. You will examine MCP's architecture, including clients, servers, and transport mechanisms, and see how MCP applications work in practice. Through guided demos and labs, you will connect to existing MCP servers and build your own MCP application.

MCP Server

In this module, you will learn how to build and enhance MCP servers. You will begin by converting tools into MCP servers and exploring simple "Hello World" examples. You will then extend server functionality with resources, prompts, and tools for real-world applications such as retrieval-augmented generation (RAG). Finally, you will explore MCP transport mechanisms, including streamable HTTP, standard IO, and deprecated SSE, while considering their security and performance trade-offs. Through guided labs, you will build and run MCP servers, connect to them using different transports, and experiment with enhanced capabilities.

MCP Hosts and Clients

In this module, you will learn how MCP clients are built and optimized for real-world use. You will examine client architecture, lifecycle management, and performance strategies such as connection pooling, caching, and load balancing. You will also explore advanced features like sampling and root controls to understand bidirectional LLM calls and filesystem boundaries. Finally, through guided labs, you will create custom MCP clients, implement advanced features, and design secure, interactive applications.