How We Work - Taylor Geospatial: For the Digital Public Good

The entire planet is imaged every day, yet machine learning for satellite imagery still lags far behind what it can do with everyday photos. That gap impedes the use of remote sensing data for necessary analysis of food security, water ecosystems, flood and drought risk, and sustainable agriculture.

Taylor Geospatial’s Innovation Program brings researchers and industry technologists side by side, turning promising academic research into open resources that serve as a foundation for what others build.

The Innovation Program

We build the bridge from research to the field.

The Taylor Geospatial Innovation Program expedites the translation of research from lab to market, bringing promising research out to the field. Each program pairs researchers with geospatial experts who drive the engineering process, and with practitioners who test results against real-world needs. Teams share and collaborate, and Taylor Geospatial’s support provides the path to try bolder ideas and change course quickly.

The cycle reinforces itself. Researchers’ work gets used more widely and becomes easier to build upon. The field gains open results and resources that serve as a foundation to advance progress. Practitioners get access to tools as the state of the art is improving.

Collaborator perspective

With the Taylor Geospatial Innovation Program, you have a really capable set of collaborators who are both technical experts and domain experts. The program creates the structure to get those people working together, and with that combination you can do remarkable things.
Caleb RobinsonPrincipal Research Scientist, Microsoft AI for Good Lab (Fields of The World Co-Investor)

Our approach to creating meaningful resources

Research to working code

Leading researchers work alongside our Technical Fellows and domain experts. Together, they deliver published, peer-reviewed research and implementations that are accessible, interoperable, and easy for geospatial practitioners to use.

Meet the Fellows

Global by design

Solving for the whole planet is a different technical problem than solving for one region. Models have to generalize across landscapes, sensors, and seasons, and data has to move at a scale few teams ever face.

Open at every layer

We want every layer to be used. Apply our global datasets as they are, or take any piece along the way: labels, training data, models, benchmarks. Researchers train on them, companies build products on them, and some teams use them for problems we never anticipated.

Everything we create is a digital public good, built to fuel discovery, open new commercial opportunities, and enable organizations addressing critical global challenges.