For many applications of machine learning, domain-specific specialists must manually determine what type of machine learning algorithms will result in the best generalization to data in the real world. To resolve this inefficiency, the Anyscale Learning for All (ALFA) group at CSAIL has begun developing Delphi, a platform to recommend predictive models for datasets. Delphi is meant to take some of the low-level human effort out of designing and implementing the best machine learning solutions to various types of data in various domains. The goal for my project will be to work to turn the first version of Delphi into a robust and useful model recommendation system.

I participated in a high-energy physics UROP involving the Higgs boson search at CERN and MIT under Christoph Paus. I have interned with several software companies, including one with a computer vision focus.