Welcome to my personal webpage!

I am a IRD Research Engineer (permanent position) appointed at the IPSL (Institut Pierre-Simon Laplace) climate modeling center located in Sorbonne Université, Paris, France. I am often referred to as an "interdisciplinary researcher/engineer", my work is focused on machine learning for climate modeling.

Before my current position, I received a PhD in Computer Vision from IMT Atlantique under the supervision of Ronan Fablet, my PhD thesis consisted in the use of analog methods (K-Nearest Neighbor Regression) to develop a new data-driven approach to tackle data assimilation: the Analog Data Assimilation (AnDA). In parallel to my thesis I investigated the use of deep learning methods for the segmentation of oceanic eddies from satellite-derived sea related maps.

I then decided to take a step towards the ocean modeling community and went to do two postdocs in two Physical Oceanography labs. First one funded by the CNES (French Space Agency) at MEOM-IGE with Julien Le Sommer on solving inverse problems for the SWOT altimeter. The second was funded by V. Balaji's "Make Our Planet Great Again" grant that I did in LOCEAN-IPSL with V. Balaji and Julie Deshayes on the tuning of climate models using surrogate modeling.

I also co-founded MoroccoAI a non-profit organization that promotes AI adoption in my birth country Morocco.

Google Scholar Profile

Researchgate

News:

I am not updating this section anymore, please connect with me on Linkedin or X for news about my work

• January 2023:
  • Our paper published in JAMES selected for an Editor Highlight in EOS magazine ! congrats to Hugo Frezat and lucky to be part of this work. Highlight
• December 2022
  • I am giving a talk for the AI4Science workshop organized by UM6P in Rabat, Morocco.
  • New paper: “Neural Fields for Fast and Scalable Interpolation of Geophysical Ocean Variables” is accepted at the Machine Learning and the Physical Sciences workshop at NeurIPS 2022. Paper; code.
• November 2022
  • New paper: "Explainable artificial intelligence for Bayesian Neural Networks: Toward trustworthy predictions of ocean dynamics". Journal of Advances in Modeling Earth Systems. paper
  • New paper: "A posteriori learning for quasi-geostrophic turbulence parametrization". Journal of Advances in Modeling Earth Systems, Paper
• September 2022
  • I gave a practical session about Physics-Informed AI for the ML4Ocean summer school. You can find my notebooks here
• May 2022
  • Good to be back to Vienna for #EGU22, I have a new presentation about History Matching for the L96 model and I am also a co-convener of the Machine learning for Earth system modeling session.
  • I was invited to give a talk at the IPSL Virtual spring school 2022
• December 2021
  • BIG NEWS, I got a permanent position at IRD as a research engineer !! I will join ESPRI and work within IPSL on climate services with a focus on collaborating with developing countries.
• July 2021
  • Spotlight talk for the workshop "Tackling Climate Change with Machine Learning" at ICML 2021 on "THOR". ClimateChangeAI
• June 2021
  • Review paper on Machine Learning for ocean modeling ! accepted in Environmental Research Letters. Paper
• May 2021
  • #vEGU2021 is online, I have two PICO presentations about History Matching for the L96 model and on our newly developed method "THOR" (collaboration with Maike Sonnewald from Princeton Univ.) and I am also a co-convener of the Machine learning for Earth system modeling session.
• February 2021
  • Webinar for the AI4ES seminars of Barcelona Computing Center link
• January 2021
  • New paper: "Physical invariance in neural networks for subgrid-scale scalar flux modeling" is accepted in Physcial Review Fluids. Paper; code
• December 2020
  • Co-chaired the AGU20 session "Innovation and Exploration in Observed and Model Oceanographic Data Using Interpretable Machine Learning"
  • Two posters in AGU20
• November 2020
  • New paper: "A Deep Framework for Eddy Detection and Tracking From Satellite Sea Surface Height Data" is accepted in IEEE Transactions on Geoscience and Remote Sensing. Paper
• October 2020
  • Poster presentation at the ECMWF-ESA Workshop on Machine Learning
• September 2020
  • Humbled to serve as a Data Challenge chair for Climate Informatics 2020 conference for the second consecutive time. If you want to discover and/or participate, we prepared a public version of the challenge on Codalab.
  • New paper: The CI2020 data challenge paper is accepted in NeurIPS 2020 workshop ClimateChangeAI. Paper;Code
• June 2020
  • "Innovation and exploration in observed and model oceanographic data using interpretable machine learning", our session proposal for #AGU20 is accepted. Please consider our session for submitting your abstracts! Link
• April 2020
  • New paper: "Filtering Internal Tides From Wide-Swath Altimeter Data Using Convolutional Neural Networks" is accepted at IEEE IGARSS 2020 conference. Arxiv; code
• March 2020
  • #EGU2020 is online this year! I have an oral presentation about tide-filtering ConvNets and I am also a co-convener of the Machine learning for Earth system modeling session.
• February 2020
  • New position: Thrilled to join LOCEAN lab at IPSL/Sorbonne Université as a postdoc researcher, will continue my work on machine learning for numerical modeling with a focus on superparametrization problems.
• October 2019
  • New paper: “Learning Generalized Quasi-Geostrophic Models Using Deep Neural Numerical Models” is accepted at the Machine Learning and the Physical Sciences workshop at NeurIPS 2019. Paper; code.
  • Pleased to serve as a Hackathon chair at Climate Informatics 2019 workshop, details
• June 2019
  • Invited talk: "Physical Oceanography meets Deep Learning"; Data Science for the future, a workshop of Global Science Week 2019, Grenoble, France.
• April 2019
  • The first IndabaXMorocco was a success! Glad to have taken part of the general chair committee.
  • Poster presentation at the European Geophysical Union 2019; Vienna, Austria. Abstract
• February 2019
  • Talk at AI & Climate-Ocean-Atmoshpere workshop; Rennes, France