Open-Source Software, Machine Learning Engineer at Mistral
France
Posted
on-site · full_time · Visa sponsorship
Apply for this role at Mistral
Responsibilities:
- Open-source model releases: package and publish models and weights to hubs (Hugging Face, GitHub) and runtimes (vLLM).
- Maintain and improve public libraries (mistral-common, mistral-finetune, mistral-inference) with tests, docs, and examples.
- Build and maintain tooling for internal research and external users (training scripts, CI, deployment helpers).
- Optimize performance and accuracy for inference and training (profiling, memory, throughput, quantization, distillation).
- Collaborate with research, engineering and the OSS community to triage issues, accept contributions, and align interfaces.
Requirements:
- Master’s degree in CS, ML, Data Science or equivalent practical experience.
- Strong Python and PyTorch skills; experience contributing to popular ML OSS projects (PyTorch, Transformers, vLLM, Hugging Face).
- Familiarity with LLM workflows: fine-tuning, supervised fine-tuning, distillation, inference optimization and MLOps.
- Experience with CI/CD, testing, packaging, and performance profiling on GPU clusters (Slurm or cloud).
- Clear communication, good open-source community practices, and ability to work across research and engineering.