Supratim Shit

My research focuses on improving the efficiency of various machine learning and deep learning algorithms by constructing coresets (data or model summarizations with provable theoretical guarantees). I am interested in theoretical and practical use cases of different coreset construction frameworks. I have a strong interest in the diverse applications of coresets within federated learning and knowledge graph domains.

I am an Assistant Professor in the Department of Computer Science at Indraprastha Institute of Information Technology Delhi. I have spent a few years as a postdoctoral fellow at Haifa, Israel. I was honored to be hosted by Prof. Dan Feldman (Haifa University) and Prof. Nir Ailon (Technion). Prior to this, I earned my PhD from the Indian Institute of Technology Gandhinagar in the Department of Computer Science. I was very fortunate to have Prof. Anirban Dasgupta as my advisor, with whom I worked on streaming coresets for various machine learning algorithms.

• COREKG: Coreset-Guided Personalized Summarization of Knowledge Graphs at IJCAI 2026

• On Sketching for Gaussian Process Regression with New Statistical Guarantees at TMLR 2026

• A Simple PTAS for Weighted k-means and Sensor Coverage at OLA 2026.