General information

Reference

2026-4001  

Position description

Category

F14 Information Systems - Engineer / Researcher

Job title

PhD in Computational Geosciences

Contract

Fixed-term contract

Contractual hours

Full-time

Contract duration

36 months

Context and contributions of the position

Joining the Bureau de Recherches Géologiques et Minières (BRGM) means becoming part of the leading public establishment in the Earth sciences, bringing together 1,000 expert and enthusiastic people in 29 entities located in mainland France and overseas.

The aim of BRGM's work is to provide geological knowledge and an understanding of phenomena linked to the soil and subsoil, with one overriding objective: to meet the challenges of environmental change through innovative projects that have a major impact on society.

Understanding subsurface fluid flow is crucial for optimizing geothermal systems and mitigating risks such as induced seismicity. Current models remain limited by the scarcity, heterogeneity, and noise of available data, as well as by incomplete knowledge of the subsurface. Physics-Informed Neural Networks (PINNs) offer a solution by integrating physical laws with complex data. However, their effectiveness is still hindered by uncertainties related to both data and models. Strengthening their robustness requires explicitly incorporating these uncertainties through "soft constraints," in order to produce more reliable and relevant simulations.

Job description

This PhD project is part of task 2.2 "Development of Artificial Intelligence methods for quantitative characterisation and forecasting of deep fluid circulation" (task leader: R. Chassagne) within the "PEPR sous-sol bien commun PC9", and will therefore use data from the Rhine Graben. The candidate will join the Geomodelling and Data Science team of the Underground Knowledge and Modelling division of BRGM - the French geological survey.

The main objective of this thesis is to contribute to the development of a comprehensive and robust model (data assimilation) of the Rhine Graben. The approach will be a hybrid data assimilation, physical model coupled with a neural network (PINNs).

The use of soft-constrained PINNs represents a major advance: it allows the importance of physical constraints to be modulated according to the quality of the data and knowledge, thus offering greater tolerance to uncertainties and gaps.

To relax these constraints and make PINNs more robust, several approaches can be considered depending on the context, each offering different ways to handle uncertainties in physical modelling:
-Weighted PINNs,
-Bayesian PINNs,
-Stochastic PINNs,
-Ensemble PINNs,
-Domain-decomposition PINNs.

Selected approaches will be tested within a dedicated data-assimilation framework. Their performance will depend on the quantity and quality of available data. The aim is a rigorous quantification of uncertainties in the final predictions.

Ultimately, the study seeks to determine how PINNs can be optimally integrated into a geothermal data-assimilation workflow.

Profile

Requested start date: 01/09/2026
Education: MSc or equivalent higher education qualification
Experience: Junior

The candidate we are looking for has:
-A good MSc degree or equivalent in applied mathematics or physics, or related area.
-Strong numerate and programming skills (Python, C/C++)
-Understanding of fluids flow in porous media
-Good communication and writing skills in English and the ability to collaborate are important.

It is a plus if you have:
-Experience in neural networks
-Experience in fluid flow modelling in porous media
-Knowledge in geothermal systems.

Working conditions

This doctoral position is available within the DCGS/GDS unit of BRGM, at the Orléans site (45).

In an environment conducive to work-life balance, particularly thanks to teleworking, BRGM stands out for its caring and friendly atmosphere.
Awarded the HRS4R label in 2021, our aim is to support you throughout your professional life at BRGM. We offer a range of schemes to help you develop your skills and expertise, with opportunities for career development and geographical mobility.
At BRGM, you will benefit from a restaurant, sports facilities, etc.

BRGM guarantees a transparent recruitment process. To apply, please send us your application (updated CV and cover letter) by April 26, 2026.
Please note that all our positions are open to everyone, at BRGM, we are committed to diversity!
We will review your application once the posting period has closed. If your application is shortlisted, you will be contacted for interviews with the hiring department, the HR department, and a cross-functional department.

Position location

Job location

France, Centre-Val de Loire, Loiret (45)

Location