HuguesKADI

Three years at the Technical Direction of EDF’s nuclear engineering division, in the Scientific IT and HPC department, scoping requirements with business clients on every project.

I built Rust monitoring tools for a petascale LUSTRE filesystem and Python log analytics, managed a cluster-access application project, and developed a Rust proof of concept for a multi-cluster abstraction layer serving 700 business users. I also led HPC and cloud workshops for our users.

Working through the INSA student-run consultancy, I deployed a state-of-the-art AI RAG (Retrieval-Augmented Generation) application in a containerised infrastructure: a business-knowledge-aware chatbot in a secure context, built on a multi-container, multi-network Docker Compose stack providing a reliable API.

I ran regular meetings with the client to sync project status and refine their needs, and supported 2 developers on their environments and technical issues.

Three months at the Scientific IT and Application Support group of École Polytechnique Fédérale de Lausanne, as a mobility within my EDF apprenticeship. Embedded with EPFL research engineers, benchmarking HPC clusters and building proofs of concept for tools like ReFrame from CSCS and Slurm-web from Rackslab.

First-level IT support for the La Poste group: receiving incidents by phone and resolving or escalating them. Very business-oriented, with a continuous stream of edge cases, and I stepped up to manage the support team for several weeks during multiple IT crises.

My first real experience of professional IT work: I ran the agency’s Windows 10 migration in summer 2019, then returned from November 2021 to February 2022 to benchmark its network storage solutions, automating the benchmarking process and building a C++ tool to compile the results.