
December 2, 2026 | 2:00 PM CEST
Live on Microsoft Teams.
On December 2, the ESA Φ-Lab Collaborative Innovation Network will host a new Φ-talk. Details are below.
Meet the speakers
Riccardo D’Ercole is an Internal Research Fellow at the European Space Agency’s Φ-lab, where his research focuses on two areas. The first is AI-driven environmental monitoring, integrating multimodal satellite and climate data with advanced deep learning techniques, for applications in vegetation monitoring, environmental forecasting, and public health. The second is Agentic AI for Earth Observation, developing autonomous multi-agent systems to orchestrate complex EO workflows, integrate heterogeneous data sources and tools, and support scientific reasoning and decision-making.
Eva Gmelich Meijling is an ESA Graduate Trainee at ESA’s Φ-lab, where she works on various applications of AI for Earth Observation (EO), focusing primarily on Agentic AI and Large Language Models, as well as Geospatial Foundation Models. Her research includes developing and benchmarking these technologies, evaluating LLMs in agentic settings, and exploring ways to connect different models, data sources, and tools. She studied Physics and Astronomy at Amsterdam University College and later Artificial Intelligence at the University of Amsterdam.
Talk abstract
Agentic AI is opening new possibilities for how AI models interact with Earth Observation data, tools, and workflows. In this talk, Eva and Riccardo will provide an overview of recent developments in Agentic AI for EO, including insights from the Agentic AI workshop organized by Φ-lab and BIFOLD, held in Berlin 19-21 October ‘26. They will then present their recent work exploring the role and potential of EO-specialized LLMs, such as EVE, within agentic frameworks, as well as PhiLiveBench, a dynamic benchmark designed to systematically evaluate LLM capabilities in agentic EO settings. The talk will conclude with future research directions, such as bridging geospatial foundation models with agentic AI and improving the explainability and trustworthiness of agentic systems for EO.
Register Here!