Staff

Function and Responsibilities:

Senior Researcher in the Global Land Monitoring group of the section Remote Sensing and GeoInformatics. (publications list)

· Conduct research, secure funding, manage projects, teams, and budgets, to develop operational tools at the interface of Earth observation, artificial intelligence, and forest monitoring, contributing to international initiatives supporting forest policy, reporting, and sustainable land management.

· Lead the development of global forest land use mapping approaches based on the FAO Forest Resources Assessment Remote Sensing Survey (FRA RSS). Designed and implemented the FLUC-16 deep learning model, trained on more than 400,000 expert-annotated FRA 2020 RSS plots, to predict forest land use classes and associated uncertainties. 

· Lead the production of a demonstration global wall-to-wall land use map (Forest, Other Wooded Land, Other Land,…) anno 2024 based on the 350,000 expert-annotated FRA 2025 RSS plots, at 10 m resolution, working closely with the FAO FRA team. This work contributes to methodological advances for future national-scale forest monitoring workflows and supports communication and visualization of FRA results.

· Lead methodological development in the ESA AI4FLUM innovation project, investigating how artificial intelligence can emulate expert interpretation of forest land use from satellite imagery while explicitly accounting for annotation uncertainty. The project is a collaboration with FAO NFO, ESA,EC JRC and Inria, and produces reusable models, open-source software, and demonstration mapping workflows implemented on cloud geospatial platforms.

· Develop and maintain open-source tools and data pipelines for large-scale geospatial analysis and AI model deployment, including the FLUC Python package for model inference on the Open Foris SEPAL platform. Design and implement machine-learning workflows integrating multi-sensor Earth observation data (Sentinel-1/2, Landsat time series, satellite embeddings) for global land-use and land-cover mapping.

· Contribute to complementary research projects linking hyperspectral Earth observation and geospatial AI, including HYPERAMPLIFAI / EnMAP downloader framework for the large-scale forest biomass estimation using hyperspectral satellite data, and the VODNet / StrucNet initiatives, which integrate satellite observations with ground-based sensors to monitor forest structure and water budget dynamics.

· Coordinate and participate in international field campaigns for forest monitoring and instrumentation deployment across Europe, Africa, and South America. Responsibilities include installation and maintenance of in-situ sensor networks (GNSS-based vegetation optical depth, LiDAR, dendrometers), terrestrial laser scanning of forest plots, data management infrastructure, and training of local collaborators.

· Disseminate results through peer-reviewed publications, open-source software, international conferences, dashboards, websites, and contribute to collaborative research networks supporting the use of Earth observation and artificial intelligence for forest monitoring and environmental policy.

Research Interests:

• Land Use and Land Cover mapping
• Forest monitoring across scales
• Software, methods and tools for reference data collection
• Data science and artificial Intelligence
• Satellite embeddings and hyperspectral imagery
  
• Instrumentation for in situ forest monitoring

Career:

• Researcher at GFZ (Potsdam, DE)
• Geospatial Consultant (Forest Resources Assessment Team) for the Food and Agriculture Organization of the United Nations (remote)
• Teacher at Le Wagon Data Science Bootcamp (Berlin, DE)
• Engineer and Scientific Liaison Officer at the Princess Elisabeth Antarctica Research Station (Brussels, BE & Antarctica).
• Doctoral Researcher at KU Leuven (Leuven, BE)

Education:

• BSc., MSc., PhD in Bioscience Engineering (KU Leuven, BE)

• Le Wagon Data Science Bootcamp (Berlin, DE)
  

Projects:

• AI4FLUM: Forest Land Use Classification and Mapping using AI models trained with the Forest Resources Assessment (FRA) Remote Sensing Survey (RSS) dataset of the Food and Agriculture Organizaiton of the United Nations (link)
• Global FRA 2025 map: first global wall-to-wall land use map based on FRA data and definitions.
• StrucNet: a global network for automated vegetation structure monitoring (link).
• VODNet: a global network for vegetation water budget monitoring (link).
• EnMAP downloader: software to download data from the hyperspectral EnMAP mission (link).
• HYPER-AMPLIFAI: forest biomass estimation using visual foundation models, hyperspectral imagery and national forest inventories (link).