Developing innovative solutions in image processing and Computer Vision, leveraging TotalEnergies’ High Performance Machine Learning (HPML) environment.

• Computer Vision Frameworks: Expertise in object detection and instance segmentation using YOLO-v8, Detectron2, and TFOD. Proficient in image processing with OpenCV scripting for R&D projects like ARGOS (robotics), MicroDeep (microalgae), Biodiversity (phototraps), and SYRACO (nannofossils).

• Workflow Automation: Streamlined and unified Computer Vision components using the Snakemake workflow management library (Python).

• Green Water Flow Quantification: Combined Sentinel-2, PlanetScope, and drone imagery with in-situ weather station data to quantify evapotranspiration for the WACAPOU R&D project. Utilized tools like Sentinel-Hub EOBrowser, ArcGIS PRO, PlanetExplorer, Google Earth Engine and QGIS.
Contributing to a poster for the SPIE Sensors + Imaging conference based in Edinburgh (September 2024)

• Biodiversity Preservation:
- Utilized the Microsoft PyTorch-Wildlife collaborative framework for automated biodiversity labeling.
- Conducted database correction and validation with the expertise of specialist ecologists.
- Performed extraction, data augmentation, and classification of animal images using YOLOv8-cls
Contributed to the preservation and census of animal species near TotalEnergies assets.

• Geological Facies Prediction: Applied Gradient Boosting Decision Trees (GBDT) models to predict geological facies using 3D seismic cubes. Evaluated I/O performance across different file systems.

• Seismic Data Analysis/Processing: Enhanced the internal DIVA seismic toolbox for SEG-Y/SEP data analysis and processing on Pangea-3/4 supercomputers.

Part of the Operational Remote Sensing Team in the Geospatial entity of DataTech during 1 year.

• Offshore Oil Slick Mapping: Specialize in mapping, interpreting, and classifying offshore oil pollution near FPSO in regions such as Africa (Gabon, Congo, Angola) and the Middle East (Qatar) using Sentinel-1A and ArcGIS PRO. Conduct monitoring every 12 days.

• Flood Mapping R&D: Contribute to the DEMETER R&D project by automating flood mapping through SAR image processing from Sentinel-1A, utilizing geemap workflows in Python and Google Earth Engine (GEE) under Google Cloud Platform (GCP). Major collaborations with the DEMETER R&D entity.

• Continuous Learning: Engage in e-learning courses on Copernicus and Mathematics for Machine Learning to stay updated with the latest advancements.

• Language Proficiency: Improving English skills through training with Intercountry, progressing from B2 to B2++ level.