Lesley De Cruz is a professor at the Department of Electronics and Informatics (ETRO), part of the ETRO.RDI Large Research Group, at Vrije Universiteit Brussel (VUB), and a senior researcher at the Royal Meteorological Institute of Belgium (RMI). She is a member of the Flanders Alliance for Climate Technology (FACT).
She obtained a PhD in Physics at Ghent University in 2011 and joined the RMI in 2012. Since 2021 she has held the joint “FED-tWIN” professorship between RMI and VUB, funded by BELSPO (Belgian Science Policy Office), combining a 50% appointment at each institution. She leads a joint VUB-RMI research group on AI and physical modelling for weather, climate, and urban applications, and coordinates the RMI’s Seamless Prediction Programme. She has authored or co-authored 34 papers in international peer-reviewed journals and contributes to open-source scientific software (MAOOAM, pysteps, mlcast).
Her research combines artificial intelligence with physical modelling across three lines.
Seamless nowcasting
The first theme is seamless nowcasting: blending high-resolution observations, numerical weather prediction, and AI for short-term precipitation forecasting. The team of Prof. De Cruz brings these scientific innovations to society by integrating them into the RMI’s short-term forecasts, and by contributing to the open-source pysteps library. The team plays an active role in the international MLCast initiative for open ML-based nowcasting. The mlcast software under development covers weather variables such as rainfall and solar irradiance, important for renewable energy forecasting.
Urban climate
The second theme is centered on urban climate and impact modelling. This covers urban heat island effects, urban hydrology, and climate-related mortality. Fast, AI-driven urban climate models enable the rapid evaluation of designs by urban planners, or the fast simulation of long-term future climate scenarios at the urban scale.
Representation learning
The third theme is representation learning for weather and climate data, including DIRESA, a dimension-reduction method for both climate data analysis and AI-based weather forecasting applications.
This research is carried out within several national and European projects (DERISC, PRECIP-PREDICT, TEMPUR, FlowCast) and international collaborations including EUMETNET and MLCast.
Prof. De Cruz is the main lecturer for Programming in Java and Techniques of Artificial Intelligence in the Master in Applied Computer Science (MACS) at VUB. She also gives guest lectures on weather and climate modelling in Physics (VUB), the Executive Master in Digital and IT Essentials (ULB/VUB), and the Postgraduate in Weather and Climate Modelling (UGent),
She initiated the LEGO Urban Climate Game, a science communication project on the urban heat island effect, and is a WATS (Women Award in Technology and Science) ambassador for Innoviris. Her work has received the EMS Outreach and Communication Award (2026) and a KVAB Science Communication Award (2025).
Watch her TEDx talk here: https://www.youtube.com/watch?v=8IraaA1qrYQ