En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)
Portail > Offres > Offre UMR7345-ERIROS-020 - Chercheur contractuel (F/H) spécialiste de l'intelligence artificielle pour prédire et contrôler le transport et les pertes de particules énergétiques dans les plasmas de fusion nucléaire (H/F)

Postdoc position (F/M) in artificial intelligence to predict and control the transport and losses of energetic particles in nuclear fusion plasmas

This offer is available in the following languages:
Français - Anglais

Date Limite Candidature : lundi 23 août 2021

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler. Les informations de votre profil complètent celles associées à chaque candidature. Afin d’augmenter votre visibilité sur notre Portail Emploi et ainsi permettre aux recruteurs de consulter votre profil candidat, vous avez la possibilité de déposer votre CV dans notre CVThèque en un clic !

Faites connaître cette offre !

General information

Reference : UMR7345-ERIROS-020
Workplace : MARSEILLE 13
Date of publication : Tuesday, July 13, 2021
Type of Contract : FTC Scientist
Contract Period : 12 months
Expected date of employment : 1 November 2021
Proportion of work : Full time
Remuneration : 2728,25 € gross mensual salary
Desired level of education : PhD
Experience required : 1 to 4 years


The successful candidate will work on the modelling of transport and losses of energetic particles in magnetic fusion plasmas, in order to increase our understanding of the underlying physical mechanisms. The first part of the work will be devoted to developing a module that will be used in the GCT particle tracking code. The second part will be devoted to the creation of a database obtained with the GCT code. Finally, the third part will be dedicated to the application of Artificial Intelligence (AI) techniques. The candidate will have to design the necessary neural architecture and train the model on the database previously built. The developed AI algorithms will have to be used to produce reduced models and to design new simulations / experiments.


The GCT (Guiding Center Tracking) code was developed at the PIIM laboratory, in collaboration with the IDRIS institute at CNRS. GCT is a guiding-centre tracking code for fusion plasmas, which allows to describe the transport and losses of thermal and energetic particles in an arbitrary geometry in the presence of a prescribed electromagnetic field.
GCT was originally developed for tokamak plasmas and very recently the experimental geometry of a stellarator was introduced. The first step in the candidate's work will be to finalize the development of the module to integrate the trajectories using the geometry of a stellarator.

Once this first step has been validated, the candidate will have to apply the GCT code to the modeling of TJ-II stellarator discharges, in order to understand the nature of the transport and losses of energetic particles in the presence of Alfvén modes.

Throughout the contract, the candidate will have to interact with our collaborators at CIEMAT (Madrid, Spain).

The database will be built using computer resources on the Jean-Zay supercomputer and the AI algorithms will be trained on the GPU server of the PIIM laboratory and also on Jean-Zay.


- Transport theory.
- Development of massively parallelized codes in Fortran 90 (MPI, OpenACC).
- Keras and Tensorflow libraries for designing and training neural networks.
- Development of pre- and post-processing diagnostics in Python.
- HPC simulations.
- Written and oral communication.
- Teamwork and international collaboration.
- English proficiency (CEFR C1)

Work Context

The PIIM laboratory has been an important player in the physics of magnetic confinement fusion machines since the 1990s. Its activities are strongly integrated at regional, national (through the Magnetic Confinement Fusion Research Federation) and European (EUROfusion) levels. The successful candidate will work under the scientific responsibility of David Zarzoso.

Constraints and risks

Screen work

We talk about it on Twitter!