Discovery Scientists, AI for Science

Mistral AI Mistral AI · AI Frontier · Paris, France · Science

Mistral AI is seeking Discovery Scientists to identify and solve key problems in STEM fields (Physics, Chemistry, Material Science) using cutting-edge AI capabilities. The role involves accelerating scientific discovery through research, improving AI models for science problems, and accelerating simulations. Responsibilities include identifying problems, leading research projects end-to-end, developing AI/ML solutions, and collaborating with multidisciplinary teams.

What you'd actually do

  1. Identify and Define Impactful Problems: Proactively identify critical challenges in your field where AI/ML can drive breakthroughs.
  2. End-to-End Research Ownership: Lead independent research projects from ideation to implementation, including literature review, hypothesis formulation, numerical experimentation, and validation.
  3. Develop and Apply AI/ML Solutions: Design, implement, and refine advanced algorithms, models, and frameworks (e.g., deep learning, LLMs, agentic AI) to address domain-specific problems.
  4. Collaborate and Innovate: Work with a multidisciplinary team of scientists, engineers, and domain experts (including customers) to push the boundaries of what’s possible with AI in science.
  5. Autonomy and Initiative: Operate with high autonomy, setting your own research agenda and driving projects to completion with minimal oversight.

Skills

Required

  • Domain Expertise in at least one scientific domain (e.g., physics, chemistry or materials science)
  • AI/ML expertise
  • Broad experience in coding for scientific computing
  • Experience with machine learning, deep learning, and/or large language models

Nice to have

  • Hands-on experience using frameworks like PyTorch or TensorFlow
  • Interdisciplinary communication skills
  • Customer Collaboration Experience

What the JD emphasized

  • first-author publications
  • Track record in AI/ML
  • further contributions

Other signals

  • AI for Science
  • deep learning
  • LLMs
  • agentic AI
  • scientific discovery