Open PhD position in Generative Models for molecular simulations, Gothenburg, Sweden

Molecular simulations allow researchers to study molecular processes in unprecedented detail. Molecular dynamics simulations have many uses, including drug discovery, materials science, and chemical reaction modeling. They can also be used to simulate biological systems such as proteins and nucleic acids. However, these simulations are computationally expensive, which limits their practical scope dramatically. In this project, the selected candidate will leverage generative AI to speed up simulations, to unlock the mechanisms of large molecular systems such as proteins.

Machine learning provides new exciting new opportunities in the natural sciences such as physics, chemistry, and biology. Its potential application areas span from designing new drugs against multi-resistant pathogens and understanding the impact of gene defects on a protein's function to speeding up computer simulations to understand fundamental scientific phenomena and design optimal algorithms for near-term quantum computers. The path toward these applications can leverage the power of deep learning to represent and process high-dimensional data effectively and encode natural laws and symmetries.

To join our group, we seek a collaborative and self-driven candidate with experience in statistical mechanics, machine learning, or dynamical systems, preferably a combination, either via coursework or through completed projects (e.g., publications or software libraries).

The Artificial Intelligence and Machine Learning for the Natural Sciences (AIMLeNS) group (head: Simon Olsson) is an interdisciplinary team of about ten researchers focusing on developing AI and ML systems to address outstanding challenges in the natural sciences. The recruited Ph.D. student will integrate into the AIMLeNS group. You can read more about the group at

The selected Ph.D. student will lead the development of generative AI systems to speed up molecular simulations. The aim is to develop a system that enables large-scale protein simulation on long-time scales, with possible applications in molecular design and structural biology. The selected candidate will build upon the group's recent work and enjoy the group's open and collaborative environment (1,2).

1) Noe, Olsson, Kohler, Wu (2019) Science 365 (6457), eaaw1147
2) Viguera Diez, Romeo Atance, Engkvist, Mercado, Olsson. ELLIS Workshop ML4Mol (2021)


Deadline for application: 29 May 2023