About
Building infrastructure through collaboration.
A computer scientist from Brittany working across high-performance computing, data logistics, and radio astronomy.
Research profile
From parallel architectures to scientific data logistics.
Mathis studies methods and tools for cross-facility workflows at IRISA and Université de Rennes. His doctoral research sits within the NumPEx French exascale program and uses ECLAT’s radio-astronomy collaborations as a demanding practical context.
His background spans engineering studies at ESIR, a one-year double-degree program at UQAC in Canada, and earlier research-engineering work on artificial intelligence, including large language models and retrieval-augmented generation.
The common thread is making complex computing infrastructure useful to people outside systems research: connecting domain scientists, software, data, and facilities without losing reproducibility or scientific meaning.
Background
Current research
PhD researcher · IRISA and Université de Rennes
Collaborative systems of systems for scientific data logistics, developed in the context of NumPEx and the ECLAT joint laboratory.
Previous research engineering
Artificial intelligence · IRISA
Engineering work on artificial-intelligence topics, including large language models and retrieval-augmented generation.
Engineering education
ESIR · Rennes
Engineering-school background complemented by a one-year double-degree program at UQAC in Canada.
International double degree
UQAC · Canada
Coursework spanning artificial intelligence, connected objects, cloud computing, and programming for parallel architectures.
Research themes
- High-performance computing
- Scientific data logistics
- Cross-facility workflows
- Distributed and federated systems
- Radio-astronomy computing
- Exascale research
- Fast I/O and massive parallelism
- HPC applications as a service
Working principles
01
Multidisciplinary by design
Computer scientists, astrophysicists, engineers, and infrastructure teams solve better problems together.
02
Complexity belongs behind a useful interface
High-level services can reduce operational friction while keeping scientific requirements visible.
03
Research should remain responsible
YoungPEx conversations include energy footprints, interdisciplinarity, and environmentally responsible research.