Research

Molecular Quantum Simulation

Classical molecular simulation approximates. We do not. By encoding molecular systems as quantum objects and simulating their evolution through variational quantum eigensolvers and tensor network methods, we recover physical behavior that mean-field approximations systematically obscure.

Our current focus is the simulation of transition metal complexes in biological contexts — systems where the spin state of a metal center determines the outcome of enzymatic reactions, and where the wrong approximation produces the wrong chemistry.

This program is conducted in collaboration with computational chemistry groups at three research universities. Publication is anticipated in the second half of 2025.

Reaction Pathway Prediction

A reaction pathway is a trajectory through a high-dimensional energy landscape. Finding the minimum-energy path — and understanding how quantum tunneling and zero-point energy modify it — requires both rigorous quantum-mechanical treatment and the capacity to explore configurations that intuition would not suggest.

We have developed a graph-neural-network architecture trained on ab initio reaction path data, capable of proposing transition state geometries and activation barriers with accuracy competitive with coupled-cluster methods at a fraction of the computational cost.

Materials Discovery

The inverse design problem — specifying target properties and recovering candidate structures — is the central challenge of computational materials science. We approach it generatively: our models propose structures, evaluate their quantum-mechanical properties, and iterate toward targets defined by application requirements.

Early-stage results in two-dimensional semiconductor design are currently under review.