Building your own AI Agent using Semantic Kernel
summary
Shows how Semantic Kernel (Microsoft .NET SDK) wires LLMs to real code: registers C# plugins/functions, enables auto-invoke kernel functions so GPT-style models call local methods, and runs agents that orchestrate up to 10 handoffs. Demos use Azure OpenAI (deployed GPT40), local Olama with Llama 3.2 (3B), and MCP servers for GitHub/Google Maps. Built a co‑pilot for a 60‑person quantum workshop and a home automation agent that unlocks a car—both working; orchestration costs measured (examples: 179, 1,400, 11,000 tokens).
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:54Semantic Kernel overview
- 3:00Setup with Azure OpenAI
- 5:55Basic chat interface
- 8:06Chat context management
- 10:35System messages for persona control
- 14:05Real-time data with plugins
- 22:16Agents for specialized tasks
- 26:53Model Context Protocol
- 31:12Custom MCP servers
- 33:44Quantum co-pilot demo
- 39:09Home automation co-pilot
- 42:18Agent orchestration
- 51:33Security with LLM
- 53:23Cost and token management
- 55:47Resources and Q&A
related talks
01h 02mDevelop AI agents with Semantic Kernel
53mUnleash the power of AI: Automate tasks with OpenAI and Semantic Kernel
46mBuilding AI solutions with Semantic Kernel
36mBuilding AI Agents in .NET with Azure AI Foundry
26mAI powered automation & multi-agent orchestration
33mHow to enhance your own AI services using Semantic Kernel
59mOrchestrate your AI with Semantic Kernel
46mAI Agents for the Rest of Us