Thank you for being here. Text Geohabari

In this episode of the Tech Stack Series, we dive into the world of GeoAI backend engineering with Brian Pondi — a geospatial machine learning researcher working on the Open Earth Monitor project, a €12.7M EU-funded initiative aimed at making Earth more observable through reusable, open-source ML tools.

Brian shares what it takes to build production-ready ML APIs that power real-world use cases like crop prediction, land cover classification, and more — with a strong focus on model sharing, discoverability, and interoperability. We explore the FAIR principles (Findable, Accessible, Interoperable, Reusable), why Europe’s research culture pushes for standardization, and how tools like OpenAPI, Stack-ML, and PyTorch shape the future of AI-driven Earth observation.

We also talk about his journey from Kenya to Germany, culture shocks, working in EU research consortia, and the importance of building tech that's open, inclusive, and reproducible.

If you're curious about the infrastructure behind the maps we use — and the AI models powering them — this is the episode for you.

🔗 Links to the Open Earth Monitor project and Brian's recommended resources:   

The Open Earth Monitor Project

Open EO