I’m a post-doctoral researcher at MSR, working in learning theory, recommender systems, and robust machine learning. Previously, I was at the University of Texas at Austin, where I graduated from with a PhD in machine learning, advised by Dr. Inderjit Dhillon.
At MSR, I work with Prateek Jain on theoretical and computational aspects of optimizing complex performance metrics arising in modern prediction tasks (such as precision at the top in multi-label learning), and active learning. I also work with Prateek Jain, Praneeth Netrapalli and Manik Varma on anomaly detection in extreme resource-constrained settings (as is the case in the internet-of-things problem space), where the detection model needs to be extremely compact, and yet achieve acceptable false alarm rates.
For a full list of publications, visit my Google scholar profile.