SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Mistral AI is seeking an Applied Scientist with deep expertise in engineering sciences to work on AI-accelerated simulation at the frontier of physics-informed machine learning. You will partner with industrial customers and internal research teams to build and deploy AI Physics Models alongside Mistral's Large Language Model offerings.
In this role, you will contribute across the full technical stack: curating high-fidelity simulation datasets from domain-specific solvers (OpenFOAM, ANSYS, COMSOL, Abaqus), training and rigorously evaluating AI models against industry validation standards, and delivering production-grade solutions directly to engineering teams. Target application domains include computational fluid dynamics, structural mechanics, semiconductor design, multi-physics modeling, and digital twins.
Key responsibilities include designing and executing large-scale simulation campaigns using industrial solvers, implementing AI model training pipelines on physics data, building automated tools for dataset creation and simulation pipeline management, developing AI agents and RAG systems that integrate LLMs with engineering workflows, collaborating with research teams to diagnose model failure modes, and managing technical projects with customer engineering teams.
You should hold a PhD or Master's degree in AI, Mechanical Engineering, Electrical Engineering, Computational Fluid Dynamics, Structural Mechanics, Semiconductor Engineering, or a related field. You must be proficient in PyTorch or JAX, write clean Python code, and be comfortable in Linux/HPC environments. Strong communication skills are essential—you'll explain complex simulation concepts to both technical and non-technical audiences. The ideal candidate is self-directed, collaborative, and has demonstrated success through industrial projects, academic research, or personal work.
Bonus qualifications include hands-on experience with simulation solvers, prior application of ML to surrogate modeling, automation of large-scale HPC campaigns, contributions to major open-source codebases, publications in top-tier venues (NeurIPS, ICLR), and a passion for code quality and testing practices.
