Open positions

Quick access:

Some open positions may not be advertised on this page, so do not hesitate to contact researchers with scientific interests close to yours directly.


  • Topic: Dependency modeling and sensitivity analysis in sedimentary dynamics
    M2 internship (6 months), with possible subsequent CIFRE PhD funding with EDF R&D. Description
    Closing date: February 2022
    Contact: Aurélien Alfonsi and Julien Reygner
    Posted in Dec 2021

PhD theses

  • Topic: Advanced numerical methods for molecular dynamics
    Description: The aim of this thesis, funded by the ERC project EMC2, will be to develop variance reduction methods in computational statistics, with applications to physical systems in molecular biology and materials science. The precise topics will be made precise in the coming months, but will in any case involve the analysis and efficient simulation of high dimensional stochastic differential equations such as the Langevin dynamics, for issues related to sampling of rare events or nonequilibrium systems. 
    Closing date: open until filled
    Supervisors: Gabriel Stoltz and Tony Lelièvre
    Contact: Gabriel Stoltz (
    Posted in Dec 2021

Postdoc positions

  • Topic: Quasistationary distributions for parallel accelerated dynamics
    Description: The aim of the postdoc will be to develop new approaches to parallel algorithms employed in molecular dynamics to accelerate the simulation of trajectories of metastable stochastic processes. It will be based on theoretical notions from the theory of stochastic processes, such as quasistationary distributions and semigroup intertwinning.
    Closing date: open until filled
    Contact: Tony Lelièvre and Julien Reygner. Do not hestitate to contact us for more information!
    Posted in Dec 2021
  • Topic: Parareal algorithms for molecular dynamics
    Description: The parareal algorithm, originally proposed in 2001, is a very efficient algorithm to solve initial-value problems (e.g. ODEs or SDEs) using parallel-in-time computations. The algorithm is based on a decomposition of the time interval into subintervals and exploits the presence of multiple processors to reduce the wall-clock time needed to obtain a solution on some time interval. The aim of the postdoc will be to develop new variants of the algorithm, better suited to the molecular dynamics context (where SDEs are considered on very long time horizons), and to test them on simple problems.
    Closing date: open until filled
    Contact: Frédéric Legoll ( and Tony Lelièvre ( Please contact us for more information!
    Posted in Jan 2022