GUMP · Gan

I initially set out with an ambitious idea: AI for Materials Science. In practice, it started quite modestly — exploring datasets on ferroelectrics and thermal management materials. Along the way, I ended up learning a great deal about LLMs. While there was no major breakthrough in materials, the journey itself turned out to be unexpectedly rewarding.

Later, I gradually built an agent system focused on thermal management materials, with the goal of accelerating material discovery, screening, and optimization. Why work on this? It's hard to give a single answer. Perhaps because research is often driven less by grand ambitions and more by simple curiosity and enjoyment.