Skip to content

Research Grants

Serenity (FNR call 2022 - granted):

“The SpacE data bRokEriNg optImizaTion sYstem (SERENITY) project aims to propose a novel type of marketplace where space data providers and consumers interact through a data broker that relies on a data lake storage. To this end, the The SERENITY project will develop novel artificial intelligence and data storage approaches that will permit the tackling of objectives like maximising providers' profits, optimising data latency, and minimising total purchase costs for customers."

I was one of the main authors of this grand. I am also now in charge of WP4, which focuses on the HPC and Quantum computing implementations of large datasets for AI/ML.

HEXAPIC (FNR call 2023 - granted):

"Plasma is difficult to characterise and model, meaning specific behaviour is difficult to predict. Since plasma consists of many particles moving in semi-order, the only correct way to describe it is to model almost every particle trajectory. This has been impossible for large systems up until now, but a new era of exascale supercomputers promises a possibility to do so. Models describing individual particle trajectories and their interactions are called kinetic models, and the best way to solve them is by using the particle-in-cell method (PIC). However, the computational task that PIC codes bring is enormous, so until now, we have been limited to simulations of small, short events."

I was the main author of this grand (from Luxembourg's side). Plus, I am now in charge of many work packages here, focusing on the heterogeneous and Quantum computing capabilities of novel plasma physics algorithms.

Quantum Computing for Scientific Computing - Kvantno računalništvo za znanstveno računalništvo (ARIS call 2025 - granted):

Addressing the climate crisis through innovative energy solutions is crucial, with fusion energy being a promising approach. The Particle-In-Cell (PIC) method, based on kinetic modeling, is widely used for both simplified and complex plasma simulations. A key challenge of PIC lies in solving large systems of equations, often handled by iterative solvers such as Krylov methods. While modern CPUs and GPUs can accelerate these computations, significant instabilities remain, highlighting the need for a unified programming model that works across different GPU vendors and compute clusters. Classical computers have long supported scientific computing and numerical modeling, especially for solving partial differential equations, but they face fundamental limitations including the memory wall, the Instruction-Level Parallelism (ILP) wall, and the power wall, which constrain further performance gains despite advances in hardware.

In light of these challenges, quantum computing emerges as a promising alternative paradigm, offering potential advantages in energy efficiency and computational speed. For instance, factoring a large number with n bits requires exponential time on classical systems (approximately \exp(n^{1/3}) ), whereas a quantum computer using Shor’s algorithm can achieve this in polynomial time (approximately O(n^2 \log n) ). Current research focuses on hybrid classical–quantum approaches to overcome limitations in simulation and modeling for fusion energy, exploring various numerical techniques and computational strategies. These efforts aim to improve stability, scalability, and efficiency in plasma simulations, ultimately contributing to effective and sustainable solutions for the climate crisis.

Research and Educational Merits