AI has moved beyond the chat window. Today's AI solutions are agents that reason, call tools, remember context, collaborate with other agents, and take action on your behalf. For developers, the question is no longer whether to build with AI, but how to build agentic apps that are reliable, observable, and production-ready. This track takes you from first principles to advanced patterns. You'll build MCP servers and agent-to-agent (A2A) solutions in C#, go from local LLMs to autonomous agents with Microsoft Agent Framework, take a practical tour of Microsoft Foundry, and give your agents memory with Azure Cosmos DB. You'll also learn what happens when the demo ends: measuring quality with eval engineering, operating agents with Aspire, building your own agent harness, and creating agents that improve themselves over time. And because doing AI responsibly matters as much as doing it at all, you'll see what goes wrong when it isn't. Want to get beyond the hype and ship real AI solutions? This track gives you the depth and pragmatic approach you need to get there.
Topics in this track include:
- Microsoft Foundry in Action: A Practical Tour of AI Agents, Knowledge, and Governance
- Building MCP Servers and A2A Agents with C# and .NET
- Building Intelligent .NET Applications: From AI Features to Agentic Systems
- Be a Better AI Developer with Eval Engineering
- Building Self-Improving AI Agents
- Building Your Own Agent Harness
- From Local LLMs to Autonomous Agents: Building AI-Native Apps with Microsoft Agent Framework
- Developing Agentic Systems in .NET: From Concept to Code
- Mastering MCP in C#: Advanced Patterns, Sampling, and “Brain Borrowing”
- Memory for Your Agents: Building Chat History and Semantic Caching with Azure Cosmos DB
- The Irresponsible AI Initiative
- When the Demo Ends: Operating .NET Agents with Aspire
- Modernize Legacy .NET Without the Big Bang
- Building Intelligent .NET Applications with AI, MCP, GitHub Copilot SDK, and Agents