Andrey Bryutkin

I am Andrey, a PhD student in Applied Mathematics at MIT, where I work at the intersection of uncertainty quantification (UQ), dynamical linear systems, and structure-exploiting Bayesian inference.

My current research focuses on:

• UQ for linear dynamical systems and their parameter reduction,
• Developing interpretable inference frameworks,
• Leveraging physical principles, such as equivariance and symmetries,
• Exploring UQ in deep learning, particularly in sampling and physical neural networks.

I am excited by questions that combine rigorous mathematics with physics-inspired models, and I enjoy designing inference and sampling methods that are not only accurate but also interpretable and structurally meaningful.

Before coming to MIT, I studied theoretical physics at ETH Zurich, where I explored representation theory in conformal field theory (CFT). I then completed my Master’s (Part III) at the University of Cambridge, specializing in probability and statistics and working on novel neural operator architectures for scientific modeling.

Outside of research, you’ll often find me on the court playing squash or tennis