I am a computational geoscientist at the University of Lausanne and a lecturer at the ETH Zurich Glaciology lab. My research lies at the intersection of high-performance computing (HPC), geophysics, and applied mathematics, with a strong focus on leveraging graphics processing units (GPUs) to accelerate scientific computing. By harnessing GPUs and supercomputers, I aim to tackle previously unsolvable problems at unprecedented resolutions.

My work centers on understanding spontaneous flow localisation in deforming porous media and ice dynamics. From a technical perspective, I develop portable, backend-agnostic HPC software and tools, enabling scalable simulations across diverse computing architectures.

I lead the GPU4GEO initiative, which focuses on developing frontier GPU-based multi-physics solvers. Recently, my research has expanded into differentiable modeling and large-scale optimisation, where we are pioneering differentiable multi-physics solvers for extreme-scale geophysical simulations using the Julia language. Our contributions integrate with the broader Julia open-source ecosystem, particularly in JuliaGPU and JuliaParallel.

As part of my commitment to democratising HPC and GPU supercomputing, I designed and launched the Solving partial differential equations in parallel on GPUs course at ETH Zurich. This course equips students in natural sciences and engineering with hands-on experience in HPC, GPU programming, and software engineering using Julia.