About Me: Computational Mathematician

I’m interested in combining modern computational and machine learning techniques and mathematics to better handle large systems of data. My current research is focused around taking medium to high dimensional data generated from dynamical systems that lie on low dimensional (unknown) manifolds and trying to numerically reverse engineer the equations that governed the original system.

In the past I worked on automatically compressing neural networks by decomposing them into smaller pieces that require fewer parameters and yet still retain the majority of the information learned by the initial network (i.e., no substantive decrease in accuracy). Even further in the past I did some research in partial differential equations on fluid dynamics with additional effects due to soap.

Please click on Demos on the top page to see various things I’ve built.