I am a PhD candidate in mechanical engineering at the University of California, Merced, in the group of Prof. Ashlie Martini. I build machine-learning models of how atoms interact, then use them to simulate materials at a scale quantum mechanics alone cannot reach.
I grew up in Bangladesh and studied mechanical engineering at BUET in Dhaka. There I ran my first molecular dynamics simulations, on how cooling rate changes the strength of the superalloy Inconel-718. That work became a first-author paper in Physical Chemistry Chemical Physics before I graduated. I also helped run a student research network that hosted 12 public webinars on computational mechanics. I still mentor students through it.
At UC Merced my work moved from classical potentials to machine-learning ones. I studied why the MoS₂ films that lubricate spacecraft fail in the cold, working with engineers at NASA's Jet Propulsion Laboratory. I tested whether Meta's universal UMA model can simulate MoS₂ doped with 25 different elements, and released the workflow as open source.
From March to August 2026 I worked at Argonne National Laboratory, training ML potentials for molten salts on the Aurora and Polaris supercomputers. I also completed the Argonne Training Program on Extreme-Scale Computing. I now continue at Argonne's Center for Nanoscale Materials, where I evaluate ML potentials for platinum–water interfaces.
I care about models that are honest about their limits. I report where a model fails, so other researchers can judge where it needs more validation or training.