PhD or postdoc position: Development of generative methods for the inverse design of materials and molecules

A position is available in Grenoble, France, to work on the development of generative models for the inverse design of materials and molecules. The position is funded by the Multidisciplinary Institute of Artificial Intelligence (https://miai.univ-grenoble-alpes.fr/) in the group of Dr Martin Uhrin with co-supervision by Prof Noël Jakse. We are happy to accept applications from candidates seeking a PhD (3 years) or a postdoc (18 months).

Project summary:  The focus of the project is on the development of cutting edge machine learning methods whose goal is to predict atomistic structures (i.e. how atoms arrange themselves) starting with a set of desired physical properties.  This will involve combining generative models (variational autoencoders, diffusion models, etc) that use invariant and equivariant representations of atomistic structures together with quantum mechanical simulations, allowing us to build algorithms that learn autonomously through active-learning, generating new data when needed.  A number of real world applications are planned, including building molecules and materials for use in next-generation batteries and for sensing.

Candidate profile: The ideal candidate will have a master's (for the PhD position) or hold a PhD (for the postdoc position) in computer science, informatics, applied mathematics or physics with experience of machine learning and an understanding of statistical methods.  Strong programming skills (in e.g. Python, C++, Julia, etc) are highly appreciated as the candidate will be expected to contribute to codes developed within the group that will be widely disseminated and used in collaborations with groups at Grenoble, EPFL (Switzerland) and MIT.

Working environment: A vibrant and highly stimulating environment that is deliberately multidisciplinary in nature, with access to collaborations at both the multidisciplinary institute in artificial intelligence (https://miai.univ-grenoble-alpes.fr/) and with physicist, chemists and materials scientists at the Materials and Processes Science and Engineering laboratory (SIMAP, https://simap.grenoble-inp.fr/) where the candidate will be hosted. The candidate will be supported to become an expert in generative machine learning methods and proficient in the generation of large databases of materials properties, both highly sought after skills for a future career in academia or industry.  They will also have access to an international network of collaborators at MIT, EPFL, Sapienza University of Rome and Microsoft and be encouraged to present their work at international venues.

How to apply: Please send your application (detailed CV, motivation letter, and names and contacts of at least two referees that can be contacted to provide recommendation letters) by email to the two advisors with the subject "MIAI ML application".  We plan to organise interviews starting mid-April.  The starting date should be before September 2024, but can be earlier depending on the candidate’s availability.  If you'd like to discuss before sending a formal application feel free to write directly to

Martin Uhrin ([email protected])
Noel Jakse ([email protected])