I was previously a machine learning scientist/engineer at General Motors. I am broadly interested in predictive modeling and problems in computational science. Prior to my position at GM, I was a postdoctoral researcher at the University of Utah. My research, primarily, focuses on developing scalable computational algorithms for design optimization and uncertainty quantification of complex engineering systems. My research interests are:

• Multifidelity-Multilevel Approaches for Uncertainty Quantification
• Numerical Linear Algebra
• Variational Inference and Statistical Learning
• Scalable Gaussian Process Regression
• Large Scale Numerical Optimization

At a personal level, I have been working on an automated trading app, transforming my previous trading ideas —rooted in chart analysis— into code. I will be sharing some of my insights that may benefit others interested in automated trading.

Please see My Automated Trading Repo

Recent News

• [May 2022] A new preprint on variational inference for nonlinear inverse problems using a neural network machinery based on hierarchical kernels, referred to as neural net kernels (NNK) Variational Inference for Nonlinear Inverse Problems via Neural Net Kernels: Comparison to Bayesian Neural Networks, Application to Topology Optimization. The codes for this work are available upon request.
• [January 2022] A new preprint is now online on a scalable GP regression approach using hierarchical matrices GP-HMAT: Scalable, O(nlog(n)) Gaussian Process Regression with Hierarchical Low-Rank Matrices. The codes for this work are available in the Codes page.
• [October 2021] I will be attending the workshop Machine Learning in Heterogeneous Porous Materials as part of Amerimech Symposium Series hosted by The National Academies of Sciences, Engineering, and Medicine.
• [July 2021] Our paper on robust topology optimization using low rank approximation and artifical neural networks is accepted in Computational Mechanics.
• [July 2021] Our paper, a joint work with Prateek Bansal and Ricardo Daziano, on using designed quadrature for maximum simulated likelihood estimation is available online: Designed quadrature to approximate integrals in maximum simulated likelihood estimation.
• [June 2021] I will be attending the 16th U.S. National Congress on Computational Mechanics. We organize a minisymposium in this conference: Robust and Verifiable Data-Driven Analysis and Design Using Machine Learning.