Summary
Pierrick Pochelu is a DevOps Engineer based in Luxembourg with a PhD and six years of experience translating HPC research into production-ready AI systems. He designed AutoML and ensemble workflows for multi‑GPU environments that produced 25% lower error at equivalent prediction time and an ensemble inference system 2.7× faster than common deployments. Comfortable from bit-level compression to distributed message passing, he has built high-performance C++/Python frameworks—including a discrete-event simulator using Petri nets and graph partitioning to enable automatic parallelism. His applied work spans patented microscopy image recognition, SAR segmentation and microgrid RL ensembles, and he regularly teaches and advises on parallel programming, HPC research and ML. Now leading tech efforts at LuxProvide on the EuroHPC Seanergys project, he combines deep optimization skills with DevOps practices to scale research prototypes into reliable production systems.
7 years of coding experience