I'm a computational structural biophysicist with a background in physics, working at the intersection of molecular dynamics simulation, machine learning, and therapeutic design. My research centers on understanding how proteins communicate allosteric signals and how those signals govern function, pursued through a combination of MD trajectory analysis, graph neural networks, and protein language models.
I'm currently developing a two-stage transfer-learning pipeline for predicting allosteric pathways and protein–protein interface hotspot residues, validated against experimentally known allosteric sites. I'm comfortable operating across the full stack of computational biophysics — from the statistical mechanics of Hamiltonian dynamics and Markov state models, to the engineering of scalable HPC pipelines and PyTorch Geometric architectures. My broader interest is in connecting atomistic simulation insight to generative design workflows, with the long-term goal of accelerating structure-dynamics-informed therapeutic discovery, including peptide, protein, and RNA design.
My path here ran through polymer and phase-separation physics, then systems biology, before arriving at machine learning for biomolecular systems — with stops at Vanderbilt, OHSU in Portland, Oregon, and now Inria in Nancy, France. Along the way I've supervised master's students at NISER and OHSU.
Download CV (PDF)Outside research
- Hiking
- Photography
- Meditation
- Reading — history, philosophy, religion, thrillers, science fiction
- Music, movies
- Teaching
- Sports — badminton & cricket