Visiting Scientist (San Francisco Office)
Core
Develop proprietary geospatial foundation models (GFMs) for satellite imagery to detect dynamic events like floods and fires.
Role type
Visiting Scientist (Postdoctoral Researcher)
Builds
Proprietary geospatial foundation models (GFMs) and time-series embeddings for PlanetScope data.
Domain
Remote sensing, satellite imagery, environmental monitoring, disaster response.
Deliverable
production ML models
Required skills
Foundation models, contrastive learning, deep learning frameworks (PyTorch/TensorFlow), Python, geospatial scientific stack (xarray, Dask, Rasterio, GeoPandas), automated data pipelines.
Preferred skills
Flood-extent mapping, water dynamics, disaster response, GFM fine-tuning (TerraMind, Prithvi, Clay), multi-sensor data fusion (PlanetScope, Landsat, Sentinel-1/2), human-in-the-loop workflows.
Technologies
PyTorch, TensorFlow, xarray, Dask, Rasterio, GeoPandas, PlanetScope, Sentinel-1, Landsat.
Responsibilities
Design and train foundation models optimized for Planet imagery with time-series integration; benchmark existing GFM architectures against PlanetScope data; build workflows for detecting short-lived events; develop methods to integrate multi-sensor data for time-series continuity; transition experimental prototypes into scalable operational products; co-author research findings for top-tier journals and conferences.
Seniority
Postdoctoral Researcher (Academic/Industry Bridge)