Loading video player...
How do we move from AI hype to real outcomes? In this talk, Vic Perdana demos SemantiClip, an open-source AI agent that transforms video into blog content using the new Agent Framework or AF (next evolution of Semantic Kernel + Autogen) and Azure AI Foundry. Learn how to build practical, interoperable agents using AF’s agent and process frameworks. With insights on orchestration patterns and when to use each framework, you’ll be ready to build real-world AI experiences This session explores the practical architecture of AI agents using Agent Framework, with a focus on composability, interoperability, and real-world orchestration. Vic will walk through SemantiClip, an open-source project that converts video content into structured blog posts using both local and cloud-based LLMs. The agent showcases AI-based planning, tool calling, and integration with the Model Context Protocol (MCP) for structured memory and cross-agent context sharing, enabling persistent, context-aware workflows. We’ll dive into AF’s agent and process frameworks, and explore how to choose between them, or use both in tandem, for different orchestration needs. Whether you’re building a co-pilot, workflow automation, or multi-agent system, you’ll leave with open-source patterns and real-world lessons you can apply right away. 🎙️ Featuring: Vic Perdana, Katie Savage and Maira Wenzel #dotnet