Aleksandr Zimin

I recently completed my PhD in Mathematics at MIT, where I was advised by Philippe Rigollet. In August 2026, I am joining the Broad Institute as a Postdoctoral Associate to work with Caroline Uhler.

I work on problems in probability, statistics, and artificial intelligence. I am interested in using computation and modern AI as tools for mathematical research: searching for proofs, testing conjectures, applying existing techniques in new settings, and finding counterexamples. I also study the mathematics of machine learning: how learning algorithms behave and how this understanding can guide the design of new models and methods.

My research includes percolation, optimal transport, economic theory, and transformer models. Recent projects include our disproof of the bunkbed conjecture, work on multistochastic and multimarginal optimal transport, applications of transport theory to auction design and taxation, and YuriiFormer, a family of Nesterov-accelerated transformers.