research

My research builds machine learning that respects scientific structure (geometry, physics, and uncertainty) and applies it across chemistry, materials, energy, and the life sciences.

My work falls into a few connected themes. Browse them by the methods I develop or by the application areas they serve; the same projects appear under whichever view fits. Each theme opens to a short overview and the papers behind it.

Geometry, symmetry & graphs

Geometry, symmetry & graphs

Equivariant graph neural networks that build symmetry into the model.
Probabilistic models & experimental design

Probabilistic models & experimental design

Gaussian and elliptical processes for uncertainty and experimental design.
Differentiable physics & scientific surrogates

Differentiable physics & scientific surrogates

Differentiable simulators and neural surrogates for inverse problems.