My primary research is in the intersection of artificial intelligence (AI) and theoretical computer science (TCS). I am in particular interested in investigating problems where AI has a possibility of making advances which are unlikely to be achieved via “pencil-paper” or traditional computational approaches. Currently, I am focussing on the following sub-domains of TCS: hardness-of-approximation, extremal combinatorics, and cryptography.

I am also interested in designing privacy preserving machine learning algorithms with strong analytical guarantees (e.g., under differential privacy), which are robust to errors in the data. My combination of academic and industrial experience has allowed me to draw non-obvious insights at the intersection of theoretical analysis and practical deployment of privacy preserving machine learning algorithms.