My research interest broadly lies in understanding and designing the self-assembly of soft matter (with particular emphasis on biomolecules) through molecular simulation. I specialize in understanding the role of solvent (e.g., water) in mediating interactions between solutes. A prime example of such interactions are hydrophobic associations, where the complex interplay between water's structural response to solutes and solute reorganization results in a myriad of natural phenomena, including protein folding and aggregation.
To drive advances in computational design of self-assembly, I actively develop novel algorithms for molecular simulation. This includes enhanced sampling and advanced free energy calculation techniques. Most recently, my work has involved the application of Variational Autoencoders (VAEs) for simultaneous molecular coarse-graining and backmapping, where models of reduced complexity and dimensionality are generated while also learning ways to reintroduce lost details. Such an ability to switch between scales is critical for multi-scale modeling, especially for maintaining thermodynamic consistency and assessing the accuracy of coarser models.
Another active area of research lies in the application of state-of-the-art machine learning and information theory tools to molecular simulation. An example of this are the VAE-based simulation techniques. My line of research focuses not only on applying these techniques, but also on integrating them on a theoretical level into statistical mechanics. Such close fundamental coupling clearly reveals how machine learning techniques can be specifically tailored or improved in order to be most useful for molecular modeling.
NRC Postdoctoral Fellowship - 2019
UCSB Graduate Opportunity Fellowship - 2018
AICHE COMSEF Graduate Student Award - 2017
NSF Graduate Research Fellowship - 2015