SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Mistral is seeking an AI Scientist to advance the frontier of AI-accelerated simulation within the AI4Engineering Science team. You will research and train foundational physics models that significantly exceed current capabilities and can be fine-tuned for diverse downstream applications by both customers and internal teams.
You will work across the full research stack: curating high-fidelity simulation datasets, designing and training novel model architectures, and rigorously evaluating them against real engineering validation standards. The role emphasizes building general-purpose foundation models that serve as the backbone for multiple downstream products rather than single-point solutions.
Key responsibilities include researching and training novel foundation models for physics simulation to push state-of-the-art in accuracy, generalization, and scale; designing and running large-scale simulation campaigns using domain-specific solvers to build high-fidelity datasets; investigating architectures and training strategies (multi-fidelity training, pretraining objectives, scaling behavior) that enable transfer and fine-tuning across diverse engineering tasks; rigorously evaluating model coverage, accuracy, and robustness against industry validation standards; and staying current with scientific developments to contribute to Mistral's frontier position in AI-for-engineering research.
Required qualifications include a PhD or Master's in CS/AI or engineering science (Mechanical Engineering, Electrical Engineering, Computational Fluid Dynamics, Structural Mechanics, EDA, Semiconductor Engineering, or related field); strong hands-on machine learning expertise with deep understanding of model architectures, training dynamics, and evaluation methodology; demonstrated experience developing ML methods for simulation or surrogate modeling; clean, readable Python code proficiency and comfort in Linux/HPC environments; fluent English with excellent communication skills; self-directed work style; collaborative, low-ego approach; and demonstrated success through industrial projects, academic work, or personal projects.
Desirable qualifications include industrial or academic experience with simulation solvers (OpenFOAM, LS-DYNA, ANSYS, COMSOL, Abaqus, Fluent, STAR-CCM+, PowerFlow, NekRS, Tau/CODA, JAX-Fluids); experience automating large-scale simulation campaigns on HPC clusters; contributions to large open-source or industry codebases; and publications in engineering or ML venues (AIAA, ASME, JFM, NeurIPS, ICLR, etc.).
