The Multimodal Vision Research Lab (MVRL) at Washington University in St. Louis is hiring a postdoctoral researcher to join a multi-institution project on next-generation neural-field representations of Earth observation data—continuous, queryable representations of the planet that can be updated as new observations arrive and that report calibrated uncertainty over user-specified regions and time ranges. Recent work on implicit neural representations, neural fields for geospatial data, and uncertainty-aware GeoAI provides the backdrop; the postdoc will help push these ideas toward systems that can be queried, updated, and trusted at scale, with validation across open-world mapping, natural hazard mapping, and ecological monitoring.
The position is funded by Taylor Geospatial and is aligned with a larger research objective toward the creation of digital public goods. Over the next two years, this effort has three key elements: hybrid neural fields, detection of features (such as buildings, agricultural fields, and wind turbines), and the creation of benchmarks for the community. The postdoc will be based in St. Louis and advised by Nathan Jacobs at WashU, with regular opportunities to interact and collaborate with a cross-institution team that includes Caleb Robinson (Microsoft AI for Good), Esther Rolf, Evan Shelhamer, Hamed Alemohammad, Hannah Kerner, Isaac Corley, Konstantin Klemmer, and Marc Rußwurm.
The position is available starting Summer 2026 (or later), with two years of guaranteed funding and the possibility of extensions. This is an in-person role on the WashU campus in St. Louis—not remote or hybrid-primary.
This is a full-time position and is benefits eligible. Salary is competitive and commensurate with experience. For additional information about benefits and other campus-level resources for postdocs, please see the WashU Office of Postdoctoral Affairs.
Submit the position application form with your CV, a short statement of how your prior work connects to neural fields and/or geospatial AI, and links to representative publications. Strong candidates will be invited to a follow-up conversation.