Microsoft Developer presents a hands-on episode with Evan Mattson, exploring how to build intelligent AI agents using Semantic Kernel, MCP Servers, Python, and Azure AI Foundry, offering practical demos for developers.

AI Agent’s Toolbox: Semantic Kernel Meets MCP and Python

Presented by Microsoft Developer, this episode is part of a series exploring the Semantic Kernel AI Agents Toolbox. Evan Mattson demonstrates how developers can leverage MCP Servers (Model Context Protocol), Semantic Kernel, Python, and Azure AI Foundry to create robust AI agents.

Topics Covered

  • Invoking MCP Tools in Agent Construction
    • Understanding MCP (Model Context Protocol)
    • How Semantic Kernel agents communicate with MCP Servers
  • Demo 1: GitHub MCP Server
    • Integrating the GitHub MCP Server using MCPStdioPlugin
  • Demo 2: Microsoft Learn MCP Server
    • Utilizing the Microsoft Learn MCP Server in your agent workflows
  • Demo 3: Azure AI Foundry Integration
    • Practical demonstration of incorporating Azure AI Foundry for enhanced agent capabilities
  • API Economy vs MCP Servers
    • Comparison and practical implications for agent developers
  • Function Calling in Agents
    • Demonstrating StreamableHTTP and plugin-based workflows
  • Getting Started
  • Community and Contribution
    • How to contribute to the open-source Semantic Kernel project
    • Connect with presenters on LinkedIn (Jorge Arteiro, Evan Mattson)
  • Watching Part 1

Key Takeaways

  • Learn how Semantic Kernel and MCP servers provide extensible frameworks for Microsoft-based AI agent development.
  • See practical implementation of agent workflows using Python and Microsoft cloud services.
  • Discover resources and sample code to accelerate building your own AI agents.
  • Understand how function calling works in modern agent architectures within the Microsoft stack.

Additional Resources

For more technical content from Microsoft Developer and Evan Mattson, subscribe to the Open at Microsoft series for regular updates.