為 約 20 位台積電同仁 進行了一場技術分享,介紹 如何使用 OpenAI API 與 Semantic Kernel 開發 AI Agent,並進一步說明 如何透過 MCP(Model Context Protocol)實現 AI Agent 的工具鏈整合與調度。透過實作與架構解析,協助學員掌握從 AI 模型調用到 Agent 行為協調的完整開發流程,為企業在 AI 應用開發上提供系統化的實踐方法。
English Description:
I delivered a technical session for around 20 employees at TSMC, introducing how to develop AI Agents using the OpenAI API and Semantic Kernel, along with integrating Model Context Protocol (MCP) to orchestrate toolchains and coordinate agent behaviors. Through practical demonstrations and architectural insights, participants learned the end-to-end process from leveraging AI models to implementing structured, enterprise-ready AI solutions.