Pranay Tankala

Ph.D. Student at Harvard CV Email

I am a Ph.D. student of Cynthia Dwork in Harvard’s Theory of Computation interested in questions about artificial intelligence related to pseudorandomness, privacy, fairness, and alignment. I was an intern on Apple’s Machine Learning Research team in 2025 and 2026, working with Parikshit Gopalan, Vitaly Feldman, and Kunal Talwar. I graduated summa cum laude from Harvard College in 2021 with an A.B. in Computer Science and Mathematics, a secondary in Statistics, and an S.M. in Computer Science, where I was on the competitive programming team.

I grew up in Cary, North Carolina and attended Cary Academy. In my free time, I enjoy playing the piano.

Pranay Tankala

Research

Structure and Randomness in AI

Many complex objects can be split into simple, orderly, and structured parts, while isolating those that are chaotic, unpredictable, or pseudorandom. This principle is exemplified by regularity lemmas in combinatorics (Szemerédi, 1975) and complexity theory (Trevisan, Tulsiani, and Vadhan, 2009). My research applies this paradigm to questions about machine learning and artificial intelligence.

Agnostic Learning by Imitating

With Sílvia Casacuberta, Parikshit Gopalan, Varun Kanade, Omer Reingold, and Konstantinos Stavropoulos
Preprint · 2026

Differential Privacy

A core tenet of responsible computing is mathematically rigorous data privacy, and differential privacy (Dwork, McSherry, Nissim, and Smith, 2006) is the leading framework for providing such protection.

Fairness and Alignment

As AI becomes increasingly involved in consequential decision-making, it is important to anticipate and navigate its pitfalls in the fairness of its treatment of a diverse user base and the alignment between its behavior and these users’ interests.

Other