Lead the Ads Recommendation team power LinkedIn Ads.

My research focuses on forming global inference from subsets of the data sources. The chosen partial data sequences are quantized (compressed) and then used to form estimates for the original data sources. We also focus on identifying the most informative data sources using limited observations in a multi-armed bandit framework. This study involves techniques from information theory, statistical learning, data compression, and sequential decision making algorithms.

Skills: Information Theory, Online learning, Sequential decision making, Data Compression, Clustering, Subset selection

Worked on scheduling algorithms in multi-user MIMO WiFi 802.11ax. Designed and analyzed new scheduling strategies and inputs for standards committee. Filed a patent based on the work.