<aside> 🐈

Junfeng Xiao

Experiments, projects, and notes from my work in AI.

AI Engineering · Model Evaluation · Agent Systems

</aside>

About

I’m Junfeng Xiao, an AI engineer focused on model evaluation and agent systems. My work has included model benchmarking, agent API integration, product research, and open-source collaboration.

This blog is where I keep track of what I’ve built and tested—the questions I started with, the problems I ran into, and what I learned along the way.

• Model evaluation: datasets, evaluation methods, API consistency, and multimodal testing.
• Agent systems: tool use, memory, planning, recovery, and prompt-injection defenses.
• AI product research: exploring how model capabilities translate into useful products.

Personal website · GitHub · Email

Explore

• Projects: projects I’ve worked on and the lessons they taught me.
• Experiments: setups, results, and approaches that didn’t work.
• Posts: technical notes, project retrospectives, and reflections along the way.

Archive

Projects

• UniFuncs Python SDK
• Hugging Face Daily Papers Digest
• Zero-Haruhi
• LLM-API-TestSuite

Experiments