Amirpasha Mozaffari
Primary tabs
Biography
I am a postdoctoral researcher in the Earth Artificial Intelligence group within the Computational Earth Sciences (CES) division at the Earth Science Department of the Barcelona Supercomputing Center (BSC-CNS).
My expertise spans machine learning (ML) applications in weather and climate, high-performance computing (HPC), workflow and data management, and Earth System Modeling (ESM). I joined BSC in 2023 and have been contributing to the optimization of climate model performance on HPC systems for projects such as RESCUE (101056939), OptimESM (101081193), and CDRESM (TED2021-130798B-I00).
Currently, my work focuses on developing ML solutions to enhance land surface representations in climate models as part of the CERISE (101082139), CONCERTO (101185000), and TerraDT (101187992) projects. Additionally, I am a Juan de Cierva Fellow (JDC2023-051208-I) with a research focus on using ML to improve weather and climate prediction.
Prior to joining BSC, I was a postdoctoral researcher at the Jülich Supercomputing Centre (JSC), where I contributed to several European and national projects, including WarmWorld (01LK2203D), IntelliAQ (787576), Maelstrom (955513), and DeepRain (01IS18047A-E).
I earned my PhD in Computational Geophysics through a joint program at Forschungszentrum Jülich and RWTH Aachen University.
My research interests include developing explainable AI for climate science, advancing workflow and data management for Earth System simulations, and improving downscaling techniques for land surface variables using ML.
Education
- PhD (Dr. rer. nat.) in Computational Geosciences, Title: "Towards 3D Crosshole GPR Full-Waveform Inversion", RWTH Aachen, Germany
- MSc in Water Resource Engineering and Management, Dissertation title: "Effect of Borehole Design on Cross-Hole Electrical Impedance Tomography (EIT) Measurements", University of Stuttgart, Germany
Research
- Workflow
- Machine Learning (ML)
- High-Performance Computing (HPC)
- Open Science
- Findable, Accessible, Interoperable, Findable (FAIR) Science