Research engineer, greenhouse gas emissions retrieval by satellite and uncertainty quantification (24 months)

Inria is the French National Institute for Research in Digital Science, of which the Inria Côte d'Azur University Center is a part. With strong expertise in computer science and applied mathematics, the research projects of the Inria Côte d'Azur University Center cover all aspects of digital science and technology and generate innovation. Based mainly in Sophia Antipolis, but also in Nice and Montpellier, it brings together 47 research teams and nine support services. It is active in the fields of artificial intelligence, data science, IT system security, robotics, network engineering, natural risk prevention, ecological transition, digital biology, computational neuroscience, health data, and more. The Inria Center at Université Côte d'Azur is a major player in terms of scientific excellence, thanks to the results it has achieved and its collaborations at both European and international level.

Job environment

This position is part of a cooperation between the Maasai team and the company QAIrbon, with the goal of improving uncertainty quantification of greenhouse gas (GHG) emission predictions. Maasai (Models and Algorithms for Artificial Intelligence) is an Inria team working on machine learning and its applications. QAIrbon is a start-up founded in 2024 developing a solution to quantify worldwide industrial CO2 emissions at the facility level leveraging satellite imagery. It aims at providing a measurement-based approach to replace voluntary reporting using emission factors prone to errors and lack of reliability.

Travel and mobility

The recruited engineer will be integrated into the Maasai team at Centre Inria at Université Côte d’Azur and will interact regularly with project partners from both Inria and QAIrbon. The position involves occasional travel for collaboration meetings between partners and participation to international conferences for dissemination of results. Travel expenses are covered by the contract.

We are seeking a highly skilled Machine Learning Researcher/Data Scientist to work on the retrieval of greenhouse gases (GHG) concentrations from satellite measurements [1,2,3], and uncertainties stemming from the interpolation of detected emissions in the concentration maps.

Satellite observations are hyperspectral images: two spatial dimensions (~30 meters resolution), and a spectral dimension (~5 nanometers resolution). A pixel is thus a vector of size N, the number of measured spectral channels (~100). The spectral dimension allows the identification of atmospheric species using the knowledge of their absorption spectrum.

Existing retrieval methods perform [4] the inversion of a state vector describing the atmospheric composition of a satellite pixel. This inversion requires many a priori estimated parameters and relies on a so-called Radiative Transfer Model (RTM) [5] as a forward model that simulates the interaction of light with the atmosphere and the ground before it is being measured by the satellite. The inversion is traditionally performed on a per-pixel basis and the RTM is a heavy full-physics model, so processing times can be long and expensive in an operational setting.

Two complementary approaches will be explored:

• Speed up and enhance the traditional retrieval framework by using neural networks as forward models. The training of such a surrogate model would be done using synthetic data generated from an existing RTM.
• Direct learning of an inverse model by a neural network to directly produce a concentration map from a satellite observation.

From these GHG concentration maps, spot emissions are detected and quantified and eventually used along other data to estimate total emissions on a continuous period of time. In this role, you will additionally apply state-of-the-art uncertainty quantification techniques—specifically Conformal Prediction [6]—to this complex, multi-stage measurement pipeline. Your mission will be to ensure our environmental data is not just highly accurate, but rigorously bounded using distribution-free confidence intervals

References

[1] C. Borger, S. Beirle, A. Butz, L. O. Scheidweiler, and T. Wagner, “High-resolution observations of NO2 and CO2 emission plumes from EnMAP satellite measurements,” Environ. Res. Lett., vol. 20, no. 4, p. 044034, Mar. 2025, doi: 10.1088/1748-9326/adc0b1. 

[2] D. H. Cusworth et al., “Quantifying Global Power Plant Carbon Dioxide Emissions With Imaging Spectroscopy,” AGU Advances, vol. 2, no. 2, p. e2020AV000350, June 2021, doi: 10.1029/2020AV000350. 

[3] M. Dogniaux  et al., “The Adaptable 4A Inversion (5AI): description and first X _CO2 retrievals from Orbiting Carbon Observatory-2 (OCO-2) observations,” Atmos. Meas. Tech., vol. 14, no. 6, pp. 4689–4706, June 2021, doi: 10.5194/amt-14-4689-2021. 

[4] C. Frankenberg, U. Platt, and T. Wagner, “Iterative maximum a posteriori (IMAP)-DOAS for retrieval of strongly absorbing trace gases: Model studies for CH4 and CO2 retrieval from near infrared spectra of SCIAMACHY onboard ENVISAT,” Atmos. Chem. Phys., 2005, https://acp.copernicus.org/articles/5/9/2005/

[5] C. Rodgers, Retrieval of atmospheric temperature and composition from remote measurements of thermal radiation, Reviews of Geophysics, vol.18, issue.7, pp.609-624, 1976, doi: 10.1029/RG014i004p00609 

[6] A. Angelopoulos, R. Foygel Barber, and S. Bates, Theoretical Foundations of Conformal Prediction, Cambridge University Press, 2025, https://arxiv.org/abs/2411.11824

Main activities (5 maximum) :

Main activities:

• Design, and implement neural architectures for surrogates of the forward model or to solve the inverse problem. Evaluate the performance of different training strategies for each task.
• Read and implement research papers that tackle similar problems.
• Compare the benefits and drawbacks of the two approaches (forward surrogate and inverse model).
• Learn, and implement Conformal Prediction using Python. Leverage modern conformal prediction techniques, first on simple datasets then for our existing GHG emission models.
• Apply uncertainty evaluation at every step of the measurement chain.

Additional activities:

• Improve conformal prediction: The project is likely to foster research questions about uncertainty quantification in general. This could lead to both empirical and theoretical work, and research papers.
• Work on other machine learning projects that involve Maasai and QAIrbon

Advanced degree (Masters or Ph.D.) in Machine Learning, Data Science, Applied Mathematics, Statistics, or a closely related field is required.

Proficiency in Python and modern machine learning libraries (PyTorch, Scikit-learn) is required.

Knowledge in spectroscopy and atmospheric science (radiative transfer modelling) is a plus.

Experience with optical satellite data (multispectral, hyperspectral) is a plus.

• Subsidized meals
• Partial reimbursement of public transport costs
• Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
• Possibility of teleworking (after 6 months of employment) and flexible organization of working hours
• Professional equipment available (videoconferencing, loan of computer equipment, etc.)
• Social, cultural and sports events and activities
• Access to vocational training
• Social security coverage

From 2692 € gross monthly (according to degree and experience)