Competitive conference systems usually require reviewers to assign a numerical rating to a subset of submitted papers assigned to them. This forms the basis of eventual acceptance of the papers. Such a system suffers from various problems, for example, a lenient reviewer is likely to assign higher scores compared to stricter reviewers; one reviewer might be assigned with poorer set of papers compared to other reviewer; a reviewers mood might end up playing a role in the rating assigned, etc. This project proposes to overcome this limitation by designing a rating system that seeks input from reviewers in form of comparisons. The project will be about designing and integrating such a system in the existing (freely available) HotCRP system.

I worked with Prof. Sandy Pentland and Yves-Alexandre de Montjoye in the Human Dynamics group at the Media Lab on information propagation in low-tech social networks. In my work with MIT startup Securigin, I reverse-engineered iOS application binaries to create security models, extract data and analyse app runtime behavior.