진학프로석·박사 채용 접수 플랫폼

Scientific environment :

The postdoctoral project is part of the new CAI4Science (AI for Science in Caen) interdisciplinary project within Caen University. CAI4Science is an ambitious research project that aims to harness the transformative potential of artificial intelligence (AI) to accelerate scientific discovery across various fields. The project brings together the expertise of four research laboratories at the University of Caen Normandie (CIMAP, CRISMAT, LPC, and GREYC) to develop and deploy advanced AI/machine learning technologies tailored to the specific needs of the domain sciences.

Scientific context :

Thermoelectric materials enable the direct conversion between heat and electricity, making them attractive for solid-state energy harvesting and cooling technologies. Thermoelectric materials are particularly appealing for applications requiring high reliability and low maintenance, such as powering sensors in IoT devices or passively cooling microcontrollers. However, their energy efficiency is constrained by the conflicting requirements of high electrical conductivity, low thermal conductivity, and a large Seebeck coefficient, properties that are rarely optimized simultaneously in most compounds (1). Addressing this challenge requires the discovery of new semiconducting materials that strike a favorable balance between electronic transport properties and have complex crystal structures that inherently result in reduced lattice thermal conductivity (2). Traditionally, finding a new material is a time-consuming task, but the progress of generative machine learning in the field of materials science has allowed us to speed up the process. Building on this momentum, we are aiming to develop new materials with optimized electrical and thermal properties for thermoelectric applications.

Scientific objectives :

The goal is to implement and apply generative deep learning algorithms to predict novel thermoelectric materials. The work will be using recent works on diffusion models for material generation (3). The postdoc will focus on the conditioning part of the generation that steers the diffusion models towards specific property values. Due to the relatively small number of materials with known thermoelectric properties, fine-tuning available diffusion models is out of reach, and the postdoc will concentrate on the inclusion of physical and chemical constraints into the generation process. At a later stage, the generated materials will be screened using numerical simulations, and promising candidates will be synthesized and characterized by chemist collaborators at CRISMAT laboratory, allowing the postdoc to focus entirely on model development and computational exploration.

(1) Complex thermoelectric materials Nature Materials2008, 7 105-114. https://doi.org/10.1038/nmat2090. (2) Enhanced High-Temperature Thermoelectric Performance of Yb₄Sb₃ via Ce/Bi Co-doping and Metallic Contact Deposition for Device Integration. ACS Applied Energy Materials2023, 6 (19), 10088-10097 https://doi.org/10.1021/acsaem.3c01693 (3) A Generative Model for Inorganic Materials Design. Nature2025, 639 (8055), 624–632. https://doi.org/10.1038/s41586-025-08628-5.

What we offer:

A stimulating and supportive research environment that fosters collaboration across disciplines.

Access to state-of-the-art computational resources.

An opportunity to contribute to cutting-edge research at the interface of artificial intelligence and materials science.

Competitive salary and benefits according to France's standards.

Support for attending national and international conferences to present research findings.

To apply:

Please submit the following documents via email:

• A CV including a list of publications.
• A cover letter outlining your research statement, background, and motivation for applying.
• A list of references to be contacted.