At Idaho National Laboratory, I develop multiscale, multiphysics, and multispecies TMAP8 models to predict tritium transport in fusion components on HPC. I use Bayesian optimization and Gaussian-process surrogate models to accelerate fusion-component design.
Previously at the University of Florida, I developed VECTOR, an open-source parallel Python toolkit for 3D microstructure generation and analysis, and contributed to PRIMME for physics-based grain-growth simulation. My work has enabled 100,000× faster preliminary fusion-plant predictions and a tritium extraction design with about 30% lower startup inventory requirements.
