Loïc Landrieu is a machine learning researcher at Imagine, ENPC, Institut Polytechnique de Paris, and an associate researcher at IGN in the LASTIG laboratory. His research focuses on computer vision and machine learning for environmental monitoring, with a strong emphasis on geospatial data, Earth observation, multimodal learning, 3D deep learning, and frugal AI.

His work explores how machine learning can help transform complex geospatial and environmental data into actionable insights for challenges such as land cover mapping, forest monitoring, biomass estimation, crop analysis, archaeology, and conservation.

Loïc Landrieu is a Hi! PARIS Advanced Fellow for the 2025-2028 period with the project “Towards Universal Representations of Geospatial Data.” This fellowship aims to develop robust and adaptable geospatial representations that can handle diverse Earth observation modalities, including hyperspectral imagery, LiDAR point clouds, radar time series, very high-resolution imagery, and historical maps.

The project focuses on multimodal self-supervised learning, robustness to spatio-temporal shifts, dataset curation, and the distillation of large geospatial models into smaller, task-specific models. Its ambition is to make geospatial analytics more accessible, efficient, and useful for environmental and societal applications.