Hugo Larcher
Machine Learning Infra Engineer at Hugging Face
Bordeaux, Nouvelle-Aquitaine, France
Summary
🤩
Rockstar🎓
Top SchoolHugo Larcher is a Machine Learning Infrastructure Engineer with 14 years of experience building high-performance computing and cloud systems, currently focused on GPU HPC clusters and infra for ML at Hugging Face. He combines a strong software-engineering background in Rust and GPU tooling with hands-on DevOps experience—implementing telemetry, nightly benchmarking, and observability for the widely used text-generation-inference project. Previously a Principal Engineer at OVHcloud, he architected cross-functional cloud, big data, and security solutions, and earlier work in aerospace fused numerical propulsion modeling with modern software stacks. Based in Bordeaux, he blends systems-level rigor from aerospace engineering with startup and product sensibilities dating back to co-founding a collaborative web startup. Notably, he bridges low-level performance work and operational reliability, making production ML workloads both fast and observable.
14 years of coding experience
12 years of employment as a software developer
Engineer's degree, Energy and Propulsion, Engineer's degree, Energy and Propulsion at INSA Rouen NormandieMaster of Science (MSc), Thermal power, aerospace propulsion option, Master of Science (MSc), Thermal power, aerospace propulsion option at Cranfield University
French, English, German
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (16)
benchmarking10
docker10
benchmark10
rust10
inference10
devops10
telemetry10
dockers10
apidoc9
api9
pytorch9
deep-learning8
nlp8
deeplearning-ai8
transformer8
Programming languages (8)
Github contributions (5)
huggingface/text-generation-inference
Jul 2023 - Mar 2025
Large Language Model Text Generation Inference
Role in this project:
Backend & DevOps Engineer
Backend & DevOps EngineerContributions:30 reviews, 17 PRs, 148 pushes in 1 year 8 months
Contributions summary:Hugo primarily contributed to the Text Generation Inference project by adding metrics documentation and API references, which enhanced the project's observability. They also implemented automatic nightly benchmarks, including the integration of Docker for running inference engines and benchmarks. Furthermore, they focused on telemetry integration by adding regular pings, simplifying error handling, and incorporating backend names and origin information. These contributions span documentation, infrastructure, and core functionality, demonstrating a focus on both backend implementation and operational considerations.
nlppytorchlanguage-modelbloombert
Hugoch/homebridge-radio
Dec 2015 - Nov 2017
Contributions:8 commits, 5 pushes, 2 branches in 1 year 10 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.