Robert T. McGibbon

I am currently a research scientist in the force field group at D.E. Shaw Research, where I work on the development of more accurate physics-based models for molecular dynamics simulations.

I earned my Ph.D. in Chemistry at Stanford University (2016), where I worked with Vijay Pande at the intersection of biophysics and machine learning on quantitative statistical analysis of protein dynamics via long-timescale molecular simulation and the development of parsimonious statistical models of these statistical data. My thesis is available here. I earned my bachelors at Princeton University (2011), where I studied Chemistry and Computer Science. While there, I worked with Andrew Bocarsly.

In my free time, I read, follow politics, and rock climb. You can contact me at

Ph.D. Research

My research at Stanford with the Folding@Home team centered around building new techniques for the quantitative statistical analysis of long timescale molecular dynamics simulations, typically for protein folding and conformational change.

This work involved both theory, software development, and applications. Aspects have involved distance metric learning, using large-margin techniques, the development of novel L1-regularized hidden Markov models for analysis of protein dynamics (including a massively accelerated multicore and CUDA implementations), and variational methods for model selection. See my publications for more information.

Most of this research required writing new software, and scalability is a central priority for our work on massive protein dynamics datasets. This work is all publicly available on github.