Agentic AI: MPC

Overview

Build secure AI agents with MCP, FastMCP, retrieval-augmented generation, multi-server workflows, and permission-based LLM interaction

What you'll learn

Describe the purpose, architecture, and operational components of the Model Context Protocol (MCP).
Configure and build MCP servers and resources using FastMCP to support secure AI agent workflows.
Create MCP clients that communicate with one or more servers through structured communication protocols.
Integrate retrieval-augmented generation (RAG) capabilities into MCP-enabled AI applications.
Analyze and apply permission models, approval workflows, and validation mechanisms for secure AI operations.
Construct and test end-to-end MCP-driven workflows for intelligent, multi-agent AI systems.

Skills you'll practise

Microsoft Certified Professional Agentic AI Responsible AI Retrieval Augmented Generation Usability AI Agents Workflow Management JSON Schema Multi-Agent Systems Artificial Intelligence Automation Large Language Modeling Auditing

Advantages and disadvantages

Advantages

University courses you can audit for free, with lectures, readings and practice quizzes.
A verified certificate from the university if you pay for it.
From IBM, a well-regarded name.
Self-paced: start any time.

Disadvantages

Graded assignments and the certificate need the paid track.
Audit access can expire a few weeks after the course ends.
Learning is free, but the certificate costs money.
Some parts (graded work, certificate) are paid.

Some points apply to every course of this kind; see how we rank.

Free to audit

Free: Choose "Audit this course" when you enrol: lectures, readings and practice are free.
Paid: Graded assignments and the verified certificate (Certificate $49). Audit access may end after the course closes.

Before you start

The following skills are required to be successful with this course: Familiarity with Python and how LLMs use tools and data is recommended. No prior MCP experience is required.