Skip to Content
View all events

Engineering AI Agents with MCP

Published by O'Reilly Media, Inc.

Learn to use the official MCP SDKs for Python, Java, and Node.js

What you’ll learn and how you can apply it

  • Build MCP servers from scratch using the official SDKs for Python, Java, and Node.js
  • Connect AI agents to your organization’s data sources, including PostgreSQL databases, local filesystems, and existing REST APIs running in production
  • Learn the Curate, Truncate, Orchestrate pattern for MCP servers
  • Scale your deployment of MCP servers using resources and custom URIs

Course description

The Model Context Protocol is the open standard supported by both Anthropic (Claude) and OpenAI (ChatGPT) that standardizes how AI agent clients connect to AI agent servers. With official SDK support across Python, Java, JavaScript, C#, and more, it has rapidly emerged as the de facto protocol for AI interoperability.

In this two-day course with Bruce Hopkins, you’ll get the complete foundation for creating real-world applications with MCP. You’ll learn about MCP architecture, core and advanced concepts, how to implement the protocol in the most popular programming languages, and how to leverage open source MCP servers to connect to your PostgreSQL database. You’ll learn how to take your organization’s REST APIs and create MCP servers with the Curate, Truncate, and Orchestrate pattern for AI agent orchestration. You’ll get hands-on experience with multiple MCP client applications such as ChatGPT and Claude Desktop and cross into the physical world to create an MCP server that lets any AI agent control real home appliances.

What you’ll learn and how you can apply it

  • Build MCP Servers from scratch using the official SDKs for Python, Java, and Node.js
  • Connect AI agents to real data sources your organization already uses, including PostgreSQL databases, local filesystems, and existing REST APIs running in production
  • Learn the CTO Pattern for MCP servers (Curate, Truncate, Orchestrate)
  • Scale your deployment of MCP Servers using Resources and custom URIs

This live event is for you because...

  • You’re a software developer or engineer who needs to connect AI capabilities to your organization’s existing APIs, databases, and internal services.
  • You have used ChatGPT or Claude for chat-based tasks, but you are ready to move beyond prompting and start building AI agents that take actions, access real data, and integrate with the systems your team already depends on.
  • You’re a technical lead or engineering manager evaluating MCP for your organization, and you need hands-on experience with the protocol.

Prerequisites

  • Python 3.10+ installed on your machine
  • Basic command-line proficiency
  • An understanding of REST APIs, HTTP APIs, and JSON
  • Basic programming experience in any major programming language like Python, Java, or JavaScript
  • User experience with ChatGPT, Claude, or other LLMs
  • An understanding of basic software development concepts (functions, variables, APIs)

Recommended preparation:

Recommended follow-up:

Schedule

The time frames are only estimates and may vary according to how the class is progressing.

Day 1: MCP Fundamentals

Introducing the Model Context Protocol (MCP) for AI agents (30 minutes)

  • Presentation: MCP architecture and core concepts; why MCP is important now
  • Q&A

MCP clients—ChatGPT, Claude Desktop, Postman, and the MCP Inspector (40 minutes)

  • Demonstration: Checking your system’s health (sysmon) with MCP
  • Q&A
  • Break

Building your first MCP server in Python (40 minutes)

  • Presentation: Setting up your project and building a weather MCP server
  • Demonstration: Using the MCP Inspector with your Python weather MCP server
  • Q&A

Building your first MCP server in Java (40 minutes)

  • Presentation: Setting up your project and building a weather MCP server
  • Demonstration: Using Postman with your Java weather MCP server
  • Q&A
  • Break

Building your first MCP server in Node.js (40 minutes)

  • Presentation: Setting up your project and building a system monitor MCP server
  • Demonstration: Using Claude Desktop with system monitor MCP server
  • Q&A
  • Break

Using the PostgreSQL database with MCP (40 minutes)

  • Presentation: Setting up and configuring the PostgreSQL MCP Server
  • Demonstration: Talking to your database in natural language without SQL
  • Q&A

Wrap-up and take-home project (10 minutes)

Day 2: Production MCP Servers

Recap (40 minutes)

  • Presentation: Take-home project results; MCP architecture and core concepts
  • Q&A

Learning the CTO pattern for creating MCP servers from existing APIs (65 minutes)

  • Presentation: Curate, truncate, orchestrate (CTO)
  • Demonstration: Creating a basic MCP server from an external API
  • Q&A
  • Break

MCP AI agent orchestration with resources and custom URIs (65 minutes)

  • Presentation: Best practices for working with multiple MCP servers; using MCP resources and custom URIs for agent-to-agent communication and data sharing
  • Demonstration: Multi-agent orchestration with resources and custom URIs
  • Q&A
  • Break

Using MCP to create AI agents for your smart home (60 minutes)

  • Presentation: Creating MCP servers for remote control of your home; creating AI agents with off-the-shelf hardware to convert dumb appliances to smart appliances without the need of hub devices
  • Demonstration: Asking a home how much power is being used by appliances
  • Q&A

Next steps (10 minutes)

Your Instructor

  • Bruce Hopkins

    Bruce Hopkins is a technical writer, an AI expert, an Intel Software Innovator for AI, and an Oracle Java Champion. He’s also the author of ChatGPT for Java and the coauthor of Beginning ChatGPT for Python.

    Xlinksearch

Skills covered

  • ASP.NET Core
  • GPT
  • gRPC
  • Web APIs