Building ML + DFT pipelines that scientists actually use

I build ML-accelerated atomistic workflows (equivariant GNNs + DFT) to discover catalysts and understand surfaces.