Research
Learning about materials, across scales.
My interests connect artificial intelligence with molecular simulation: how we describe interactions, represent dynamics, and study material properties.
AI for Materials
Machine learning for molecular and materials modeling, with an interest in the relationships between structure, dynamics, and macroscopic properties.
- AI for Science
- Materials modeling
Molecular Dynamics & Machine Learning Potentials
Learned interatomic interactions for molecular simulation, with interests in the application and computational efficiency of machine learning potentials.
- Molecular dynamics
- ML interatomic potentials
HPC Optimization of ML Interatomic Potentials
I am particularly interested in where MLIPs spend computation and GPU memory, and whether redundant work can be reduced while retaining useful physical predictions. Improving memory efficiency and simulation throughput is a research direction I want to pursue; I do not yet report optimization or training results.
- GPU memory
- Computational efficiency
- HPC
Coarse-Grained Molecular Modeling
Reduced molecular descriptions that connect atomistic detail with modeling at larger length and time scales, and the information those descriptions retain about molecular dynamics.
- Coarse-graining
- Multiscale modeling
Representation Learning / JEPA
JEPA-style representation learning for molecular dynamics and coarse-grained modeling, with an interest in representations that capture molecular structure and its evolution over time.
- Representation learning
- Molecular dynamics
Polymer Property Prediction
Machine learning and molecular simulation for studying polymer properties, including glass-transition temperature and density. I am also interested in agentic workflows that support polymer research.
- Polymers
- Property prediction
- Research workflows