Praxis AI

After a few projects, my partners and I wanted to find out whether RAG and AI agent technology could become a real product. That became Praxis AI (喬喜科技). I am a co-founder, responsible for the AI and RAG technical direction, system planning and exploring how companies would use it.

We were later accepted into the NTU Garage incubator at National Taiwan University.

The biggest lesson: being corrected by reality

At first we were very technology-driven. We assumed that because we had the technology, companies would need it. Then the team went to interview managers at a factory and heard front-line people describe what actually bothers them every day.

After that interview we went back and re-examined our product assumptions, and started shifting from what we can build to what they actually need.

What I learned

This was the first time I looked at things from outside system development: product requirements, usage scenarios and how technology lands in practice. Between “it works” and “people are willing to use it” there are still many design and system challenges. Honestly, that road is much harder than writing code.