You will be in charge of open-sourcing state-of-the-art models, whilst maintaining and improving Mistral’s publicly available libraries. Your work is critical in helping turn research breakthroughs into tangible solutions and improve Mistral's open-source ecosystem.
About the Open Source Software team
Our OSS team is embedded in our Science team and works very closely with various engineering and marketing teams. All OSS team members can fluidly move on the production / research spectrum depending on where the needs are or where their interests lie
Responsibilities
• Releasing our models to open-source platforms and libraries, e.g., vLLM, GitHub, Hugging Face
• Maintaining Mistral’s open-source libraries (mistral-common, mistral-finetune, mistral-inference)
• Create and maintain tooling and services: both internal facing (internal research) and external facing (open-source libraries)
• Implement and optimize open-source and internal libraries for performance and accuracy, ensuring production readiness and employing cutting-edge technology and innovative approaches
• Collaborate with the open-source community (PyTorch, vLLM, Hugging Face)
About you
• Master’s degree in Computer Science, Machine Learning, Data Science, or a related field
• Experience contributing to popular open-source libraries such as PyTorch, Tensorflow, JAX, vLLM, Transformers, Llama.cpp, …
• Passion for contributing to the open-source software ecosystem
• Expert programming skills in Python, PyTorch, MLOps
• Adaptable, proactive, and autonomous
• Attention to detail and a drive to go the last mile to build almost perfect tools
• Deep understanding of machine learning approaches, especially LLMs and algorithms
• Low-ego, collaborative and have a real team player mindset
Now, it would be ideal if you have:
• Experience with training and fine-tuning large language models (e.g., distillation, supervised fine-tuning, policy optimization)
• Experience working with Slurm
• Worked with research teams before
• Experience as a core-maintainer of a popular ML open-source library
XML job scraping automation by YubHub
