P. Bernát Szabó

About

I am a senior researcher in quantum chemistry in the AI for Science lab of Microsoft Research, working on generating accurate training and validation data for the Skala density functional.

I was introduced to quantum chemistry in the group of Prof. Mihály Kállay (opens in new tab) where I developed the open-shell coupled-cluster code of MRCC (opens in new tab). With supervision from Péter Nagy (opens in new tab), I also extended this implementation with local correlation approximations (opens in new tab) to exploit the short-ranged nature of electron correlation. Later on, I worked with both Prof. Arnout Ceulemans on the solutions of Jahn-Teller problems (opens in new tab) and Prof. Liviu Chibotaru on the description of the lanthanide luminescence (opens in new tab). My Master’s thesis focuses on using cosolvent methods to discover cryptic binding pockets (opens in new tab) of proteins.

I wrote my doctoral thesis on variational quantum Monte Carlo methods with deep-learning ansatzes (opens in new tab) under the supervision of Prof. Frank Noé (opens in new tab). Using the developed methods, we obtained barrier heights (opens in new tab), electronic excitation energies (opens in new tab), and electron densities (opens in new tab) of strongly correlated molecules at unprecedented accuracies.

Since January 2026 I am a full-time senior researcher in the AI for Science organization.