Hi, I'm Niyathi 馃憢

nk6074@princeton.edu github.com/niyathikukkapalli linkedin writing

I'm an undergraduate studying mathematics and computer science at Princeton University. My research sits between machine learning and theoretical computer science: how structured, mathematically generated data can train better models, how neural networks can learn to reason about strategic and formal systems, and how geometric algorithms scale beyond Euclidean space.

In the Princeton Vision & Learning Lab, I work on procedural generation of synthetic datasets for computer vision, where every generated scene comes with exact ground-truth labels. At the University of Maryland, I work in Sarah Miller's group on deep learning architectures for predicting outcomes in game-theoretic settings, with applications to formal languages. With Donald Sheehy's group at NC State, I worked on range search over products of metric spaces with bounded doubling dimension; this work was accepted to CCCG 2026.

On the practical side, I'm a technical lead at TigerApps, where I manage a team of student developers and review and merge pull requests for apps used across campus, and I build full-stack products and backend infrastructure, most recently Nimzo. I'm also a course assistant for COS 324, Princeton's introduction to machine learning, and a peer academic advisor.

More broadly, I'm interested in AI agents, generative AI, theoretical machine learning, probability theory, topology, and measure theory.