Postdoctoral researcher in differentiable modelling for extreme-scale geophysics simulations

Introduction

UNIL is a leading international teaching and research institution, with over 5,000employees and 17,000 students split between its Dorigny campus, CHUV, and Epalinges. As an employer, UNIL encourages excellence, individual recognition, and responsibility.

www.unil.ch/carrieres

Presentation

The Institute of Earth Sciences (ISTE) invites applications for a postdoctoral position within the ∂GPU4GEO project ( https://gpu4geo.org/dgpu4geo/ ). 

This role offers the opportunity to contribute to the next generation of computational geoscience : developing differentiable multi-physics solvers in Julia for extreme-scale geophysics simulations. The successful candidate will work at the forefront of high-performance computing, combining advances in automatic differentiation, numerical methods, and GPU-accelerated architectures to design and run cutting-edge large-scale simulations.

You will collaborate with an international network of researchers, HPC experts, and domain scientists, including teams at the Swiss National Supercomputing Center (CSCS), thePronto Lab (University of Illinois), the Geophysical Fluid Dynamics Group (ETH Zurich), and the University of Mainz, amongst others.

Job information

Start date : 01.01.2026
Contract : 2 years
Work rate : 100% research
Location : Institute of Earth Sciences, Faculty of Geosciences and Environment, University of Lausanne

Main tasks

• Develop differentiable, MPI-parallel, GPU-accelerated Julia software for nonlinear multi-physics problems on staggered grids using the accelerated pseudo-transient method.


• Implement and optimize codes for modern GPU supercomputers (NVIDIA and AMD), including ALPS (CSCS) and LUMI (CSC).


• Design large-scale Julia applications, including parallel data handling, I/O, and advanced in-situ visualization.


• Produce a flagship large-scale simulation within the project.


• Collaborate closely with a second postdoctoral researcher working on complementary topics.

Your profile

Required qualifications :

• PhD in numerical geology, geophysics, computational science, or a related field.


• Strong interest in automatic differentiation, scientific computing, and scalable multi-physics solvers.

Beneficial experience :

• GPU development in Julia (CUDA.jl for NVIDIA and AMDGPU.jl for AMD).


• Parallel programming (especially MPI–MPI.jl).


• Software development in Julia.


• Modern collaborative software practices (Git, CI/CD, testing).

What we offer

• A unique opportunity to advance differentiable multi-physics modelling in Julia for extreme-scale geophysics applications.


• Professional growth and leadership opportunities within a collaborative andinternational research environment.


• A dynamic, multicultural, and supportive workplace at ISTE, with training and networking opportunities.


• Work conducted primarily in English ; knowledge of French is not required.

More information on  www.unil.ch/carrieres

Contact for further information

Dr. Ludovic Räss (E-Mail schreiben)

Application

Deadline : 01.10.2025

Please submit a single PDF file containing :

• Cover letter (motivation and fit for the position).


• Research statement (max. 1 page).


• CV (max. 2 pages).


• Copy of PhD diploma (or confirmation of successful defence).


• Contact details of two referees.

Applications must be submitted via the UNIL online system.

Additional information

UNIL is committed to equal opportunities and diversity:
https://www.unil.ch/egalite/en/home.html
UNIL provides welcome services for new arrivals:
https://www.unil.ch/welcomecentre/en/home/menuguid/services-daccueil-unil.html
Information on services provided to families at UNIL can be found here:
https://www.unil.ch/familles/en/home.html