SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Mistral is building full-stack AI solutions—from frontier models to developer tools and compute infrastructure. As a Research Software Engineer on the software side, you will design and harden the codebase, tools, and distributed services that enable scientists to train and ship frontier-scale models. You do not need prior ML experience; what matters is writing clean, reliable, scalable code.
You will join the Research Software Engineering team (based in Paris and London) to build and maintain shared dev-tools, evaluation and data pipelines, training frameworks, cluster tooling, and CI/CD infrastructure. Your responsibilities include:
• Accelerate researchers by owning complex parts of large-scale pipelines and delivering robust internal tooling.
• Interface research with product: expose clean APIs, automate model pushes, surface live metrics.
• Write efficient, well-tested Python and systems code; enforce code review, CI, and observability standards.
• Design and optimize distributed services (Kubernetes, SLURM, managing thousands-of-GPU jobs).
• Prototype utilities (CLIs, dashboards) and carry them through to stable, shared libraries.
You should have a Master's in Computer Science (or equivalent), 4+ years building and operating large-scale or distributed systems, strong software-design instincts (modular code, tests, CI/CD, observability), fluency in Python plus one systems language (C++, Rust, Go, or Java), and hands-on experience with container orchestration and schedulers (Kubernetes, SLURM, or similar). Comfortable profiling performance, optimizing I/O, and automating workflows. Self-starter, low-ego, collaborative, high-energy.
Nice-to-haves: exposure to ML workloads or data-processing pipelines, GPU cluster or CUDA experience, open-source contributions or widely used internal tools.
The role is hybrid, primarily based at one of Mistral's European offices (Paris, Warsaw, Zurich, London). Remote candidates based in France or UK are considered, with requirement to visit the local office for the first week of onboarding and at least 3 days per month thereafter.
