Beyond the Sandbox: LLMs for Planet-Scale Remote Sensing

PowerTalk

About the session

As Large Language Models (LLMs) evolve, they are shifting from coding assistants to active research partners in Earth observation. This talk explores how LLMs are transforming environmental AI through collaborative development and automated scientific discovery. We examine how treating LLMs as "co-thinkers" accelerates complex pipelines via human-AI peer-coding and introduce the Empirical Research Assistant (ERA), a Gemini-powered system that uses automated, closed-loop iteration to refine algorithms for tasks like CO2 estimation. Furthermore, we showcase how AI co-pilot, Earth AI and similar tools can be utilized for fine-grained remote sensing tasks to aid climate mitigation and adaptation efforts.

Speaker

Download Brochure