Research Scientist, Biosphere Models

We seek a creative researcher with strong Geospatial AI and Earth Observation expertise to design, implement, and train state-of-the-art models on planetary-scale datasets. You will pioneer self-supervised and weakly supervised pretraining methods to address data scarcity in natural environments. You will build scalable data pipelines for ingesting heterogeneous Earth Observation data and lead the technical validation of models against ground truth. You will collaborate with domain experts to refine model objectives for downstream application domains and design robust validation campaigns. You will communicate findings clearly and contribute to open-source releases and scientific publications, influencing both research and product decisions. A track record in GeoAI, EO data modalities (vision, SAR, LiDAR), and experience with distributed training and geospatial data processing tools is highly valued.