Advisors: Klaus Schulten (Theoretical and Computational Biophysics Group), Martin Gruebele and Taras Pogorelov Computational modeling/simulation, machine learning, and statistical analysis of protein dynamics and function:• Conducted molecular dynamics simulations with supercomputers utilizing parallel computing code to study biological processes (protein folding, proteasome recognition, microtubule dynamics, etc.).• Gathered terabyte-sized time series trajectories and analyzed them with unsupervised learning, matrix factorization, clustering, and Markov state models, to extract patterns from protein structures.• Designed correlation data analysis method to gain insights in protein folding rate.• Implemented a pipeline of code to gather data from experimental outputs and to validate with simulation results.• Collaborated as a theorist with leading experimental scientists of diverse backgrounds, providing insights that guided or confirmed their experiments.• Contributed to a successful multi-million-dollar grant proposal to the National Institutes of Health.

Partnered with a startup on building predictive models for sensor time series:• Consulted for a startup company that builds sensors to track punches for fighters.• Built machine learning models (logistic regression, neural network, and random forest) from 20k punch time series to provide actionable insights into product development.• Enhanced algorithm to differentiate punch vs. non-punch at 99% accuracy, and to precisely predict punch begin point and end point with error range of ±89 ms and ±14 ms, respectively.

• Examined spatial distribution of plant seed banks and performed statistical regression analysis (LASSO) to understand competition between native and invasive species.

• Investigated rhodopsin activation in membrane bilayers through simulation and docking methods.