Geospatial Foundation Models
Supervisor: David Coomes
Project description
Geospatial foundation models such as Tessera and AlphaEarth leverage petabytes of freely available remote sensing data to generate embeddings that can be efficiently applied to tasks like land use classification.
These self-supervised approaches enable analyses at unprecedented spatial scales – capabilities that were once limited to large corporations or government research agencies. The main current constraint lies in the availability of labelled datasets for downstream applications.
Through partnerships providing access to extensive field datasets, this project aims to produce new, continent-scale maps of habitats and temporal changes. Such work holds significant value for practitioners and policymakers striving to manage the planet’s resources more sustainably.
Working environment
The role holder will work with teams from across the research group, the Centre for Landscape Regeneration and other disciplines, with lots of opportunity to learn and network. Some flexible working can be accommodated, but the role holder will need to be in the office in Cambridge regularly.
Essential knowledge, skills and attributes
- You should be studying, or have completed a degree in a relevant subject such as environmental sciences, physical sciences, engineering, computer sciences or similar
- Applicants would benefit from a knowledge of large-scale data analysis, nature based solutions and environmental issues
- Experience and skills in use of Python, with experience in R also being beneficial
- An interest in geospatial analysis, geospatial foundation models and/or computer vision
- Good analytical skills; able to collate, understand and draw conclusions from quantitative and qualitative information
- Good interpersonal skills; confident in meetings with both internal and external colleagues and working with people from a range of backgrounds and disciplines
- Well-developed organisational and timekeeping skills and the ability to manage projects
- Ability to prioritise own workload and plan effectively as part of a small team.
Background information
This project is part of the Centre for Landscape Regeneration, led by David Coomes. He is currently recruiting a new PDRA who will also be involved in the supervision of this intern.
Contact us
If you have questions about this project, email David Coomes