Welcome to my GitHub! I'm Pan Jiaxiang, a graduate student interested in LLM, AI Agents, and post-training.
I enjoy building reliable Agent systems, experimenting with model behavior, and turning research ideas into reproducible engineering projects.
- 馃摍 Focus: LLM 路 AI Agent 路 Post-training 路 Tool Use 路 Evaluation
- 馃敤 Creator: Building a general-purpose Agent Harness with Runtime, Tool Calling, Trace/Replay, permission control, and reproducible evaluation
- 馃敩 Research: Exploring LLM/Agent reliability and post-training, with previous experience in 3D reconstruction, reinforcement learning, and computer vision
- 馃摑 Sharing: I enjoy organizing and sharing what I learn about AI, research, and engineering
- 馃崠 Fun Fact: Meat lover 馃崠


