Turning hard science into usable intelligence

Nikita Kazeev is a Research Fellow in the laboratory of Professor Kostya Novoselov at the National University of Singapore (NUS). His work starts from the premise that, if you look deep enough, most of science is approximate computation, and machine learning is the best tool we have for building those approximations – a way to expand the frontier of what we can predict and, ultimately, control in the physical world.

He holds a dual PhD in Computer Science from HSE University and in Physics from Sapienza Università di Roma, studied as an undergraduate at the Moscow Institute of Physics and Technology, and graduated from the Yandex School of Data Analysis.

Prior to joining NUS, Dr. Kazeev worked at CERN, applying machine learning to structured data and ML-driven simulation in high-energy physics; he received the 2025 Breakthrough Prize in Fundamental Physics as a part of the LHCb collaboration.

At NUS his research centers on making AI an instrument for scientific discovery – starting with the design of new materials, and reaching beyond. He is deeply involved in building the AI-for-science community: he was the main organizer of the ICLR 2025 Workshop on Machine Learning for Multiscale Processes and Program Chair for the AI4X 2025/6 conference, and is co-Principal Investigator on a US$3.4 million AI Singapore grant for multiscale machine learning. He works on turning AI into a fair and objective tool for science governance.

Beyond academia, he is an AI Advisor at Perfomax and has been a machine learning consultant to the Constructor Group since 2021, helping translate frontier methods into shipped products across science and industry.