I am an assistant professor in the Department of Computer Sciences, at the College of Computing and Artificial Intelligence at UW-Madison. I am also affiliated with the Department of Statistics. I work on theoretical topics in machine learning and game theory. For more details, see my publications. Prior to UW-Madison, I was at UC Berkeley, Carnegie Mellon University, and the University of Moratuwa, Sri Lanka. See here for a formal bio.

Selected publications: Classical learning theory measures the complexity of learning by the number of samples needed, implicitly treating samples as interchangeable units that cost the same. In practice, the cost of a data point or experiment varies with what is being collected and from whom. Our work, supported in part by an NSF CAREER award, develops the statistical and game-theoretic foundations of cost-aware learning. This includes deciding how much data to gather from heterogeneous sources, sequential optimization where each query has its own cost, and designing incentives so that strategic data sources collect and share data truthfully and fairly. We develop policies for data collection, adaptive optimization, data sharing, federated learning, and data marketplaces, with provable guarantees on performance, cost, incentive compatibility, and fairness. Below are some relevant papers:

• Mechanism Design for Collaborative Normal Mean Estimation [arxiv],   NeurIPS 2023
• Learning to Price Homogeneous Data [arxiv],   NeurIPS 2024
• Collaborative Mean Estimation Among Heterogeneous Strategic Agents: Individual Rationality, Fairness, and Truthful Contribution [arxiv],   ICML 2025
• Balancing Performance and Costs in Best Arm Identification [arxiv],   NeurIPS 2025
• A Cramér-von Mises Approach to Incentivizing Truthful Data Sharing [arxiv],   NeurIPS 2025
• Pairwise Exchanges of Freely Replicable Goods with Negative Externalities [arxiv]
• Learning from Biased and Costly Data Sources: Minimax-optimal Data Collection under a Budget [arxiv],   COLT 2026
• Fair Division of Work in Collaborative Mean Estimation via Bargaining [pdf coming soon],   NeurIPS 2026

Openings: I am recruiting students with strong backgrounds in mathematics and statistics. Please read this before emailing me.

Current teaching (Fall '26):   CS540 - Introduction to Artificial Intelligence

PhD students

Alex Clinton (CS)
Michael Harding (Statistics)
Joon Suk Huh (CS)
Peter Yang (CS)

Undergraduate students

Ishaan Kharbanda

Contact

Morgridge Hall 5506
University of Wisconsin-Madison
Madison, WI 53706.
email:   kandasamy [at] cs (dot) wisc {dot} edu     (Please read this before emailing me)