AI Agent Exploration: Multi-Role Task Processing Agent
- Time
- 2026-08-08 13:00 ~ 13:30
- Speaker
- Yale Hsu
- Room
- TR311
- Co-write
Abstract
- Common Terminology in AI Agents
- Why I Decided to Build My Own Agent Harness
- How I Started This Project
- The Key Building Blocks of an AI Agent
- External Systems I Integrated
- Demo
About Yagent Yagent began as an experimental project that has since evolved into a Multi-Role Agent I use for light daily tasks. Initially, I started this project to explore the world of AI Agents and understand how they work under the hood. As coding agents got increasingly powerful, I began building out my own custom agent framework while trying it out myself.
While it’s not yet fully mature, I’m continuously iterating on it to gradually migrate my own workflows onto the system. So far, it’s steadily moving toward that goal.
