Microsoft Developer presents an in-depth exploration of building and orchestrating AI-powered multi-agent systems with Microsoft Foundry, highlighting advanced frameworks and deployment practices for scalable solutions.

Ignite BRK197: AI Powered Automation & Multi-Agent Orchestration in Foundry

Presented by Microsoft Developer

Overview

This session introduces developers to advanced methods for constructing and scaling multi-agent systems with Microsoft Foundry. It emphasizes moving beyond single-agent prototypes towards orchestrating a ‘fleet’ of AI agents using Microsoft technologies.

Key Technologies & Tools

  • Microsoft Foundry Agent Framework (combining Semantic Kernel and AutoGen)
  • Shared State Management
  • Human-in-the-Loop Integration
  • OpenTel and MCP Toolchains
  • Agent-to-Agent (A2A) Communication and Activity Protocol
  • LangGraph and OpenAI Agents SDK Framework Support

Multi-Agent System Building Blocks

  • Use Semantic Kernel and AutoGen for agent logic and orchestration
  • Implement shared state for agent collaboration and context persistence
  • Integrate human stakeholders for oversight and hybrid workflows
  • Utilize OpenTel tools for agent observability and monitoring
  • MCP toolchains support robust development and deployment
  • Deploy agents using governance and containerization best practices in Microsoft Foundry

Extensibility and Deployment

  • Bring in popular external frameworks: LangGraph, OpenAI Agents SDK
  • Deploy scalable, governed, and observable agents in production on Microsoft Foundry
  • Leverage Azure for underlying infrastructure and cloud scalability

Resources & Community

Conclusion

Microsoft Foundry provides a robust foundation for building, orchestrating, and scaling AI multi-agent systems within the Azure ecosystem. Developers can leverage an array of tools and frameworks to create advanced, production-ready AI solutions.