At LDEO Columbia University, I develop machine-learning models for ice-sheet and climate research. My projects include MAR-IA, an XGBoost emulator for daily Greenland meltwater and surface mass balance prediction.
I also developed neural-network models that map surface, bed, and geophysical observations to basal friction, then evaluated predicted friction fields in ice-flow simulations and tested transferability in West Antarctica.
At Geminus.AI, I developed physics-informed surrogate models for industrial process simulation, design optimization, and transient PDE prediction. Deployed workflows reduced runtime from hours to seconds, and I created root-cause analysis methods for multi-component digital twins.
My earlier work included CFD optimization at Mitsubishi Electric Research Laboratories and machine-learning-augmented turbulence modeling at Iowa State University. I hold a PhD in Aerospace Engineering from Iowa State University.
