Currently working on modeling 5-dimensional plasma turbulence in Tokamaks (gyrokinetic simulations) with machine learning, a joint project with the UK Atomic Energy Authority, as well as domain adaptation and robustness for industrial processing in a variety of application and manufacturing conditions in collaboration with the Linz Center of Mechatronics.

During my Master's, I applied Graph Neural Networks and Equivariance to particle fluid problems. In the past I was involved in an autonomous racing student club, as well as reinforcement learning and computer vision projects.
I've also worked in industry throughout my Master's, on process mining and simulation at Celonis.

If you are looking for a BSc/MSc thesis in my research areas, reach out at galletti[at]ml.jku.at

SIMSHIFT: A Benchmark for Adapting Neural Surrogates to Distribution Shifts -- [code]
P. Setinek, G. Galletti, T. Gross, D. Schnürer, J. Brandstetter, W. Zellinger
↪ 2025, preprint.

5D Neural Surrogates for Nonlinear Gyrokinetic Simulations of Plasma Turbulence
G Galletti, F Paischer, P Setinek, W Hornsby, L Zanisi, N Carey, S Pamela, J Brandstetter
↪ 2025, ICLR Workshops (CCAI & MLMP).

LagrangeBench: A Lagrangian Fluid Mechanics Benchmarking Suite -- [code]
AP Toshev*, G Galletti*, F Fritz, S Adami, NA Adams
↪ 2023, NeurIPS Datasets and Benchmarks.

Learning Lagrangian Fluid Mechanics with E(3)-Equivariant Graph Neural Networks -- [code]
AP Toshev, G Galletti, J Brandstetter, S Adami, NA Adams
↪ 2023, Geometric Science of Information.

Last update April 2025

Template by Jon Barron.