AI Scientist - Agentic Engineering
Skills & stack
About the role
The AI Scientist, Agentic Engineering position at Mistral AI is a middle-to-senior role based in Paris, France. You will join the AI4Engineering Science team to expand agent capabilities across engineering lifecycle domains, including CAE and semiconductor design.
Work format
This is a full-time position with a hybrid work arrangement. The stated English level is C1.
Skills and tools
- Python
- LLM
- Reinforcement Learning
- Linux
- HPC
What to prepare
- Prepare a walkthrough of an engineering workflow you have run, including how you assessed task correctness.
- Choose a Python code sample that demonstrates readability in machine learning work.
- Prepare an example of diagnosing an LLM failure, covering the evaluation method and findings.
Responsibilities
- Design pretraining, supervised fine-tuning and reinforcement learning data for CAD, CAE and semiconductor/EDA tasks.
- Define verifiers and evaluation criteria that assess whether engineering tasks are completed correctly.
- Design and improve agent architectures for long-horizon engineering workflows.
- Build evaluation benchmarks and diagnostic tools to identify model failures.
- Work with domain experts to identify high-value engineering workflows.
- Contribute to pre- and post-training research across the science organization.
Requirements
- Hands-on machine learning expertise, with depth in LLM development and reinforcement learning.
- Demonstrated experience running engineering workflows in CAD, CAE, semiconductor simulation or EDA.
- Write clean, readable Python code and work in Linux and HPC environments.
- Communicate clearly, collaborate with others and direct your own work.
- Willingness to learn at the intersection of machine learning and engineering.
Apply for this role
Send a short intro — the hiring team will review and respond.
Paris-headquartered AI company building open-weight and commercial large language models and an enterprise AI platform (Le Chat, La Plateforme). Incorporated in Paris, 900+ employees, with science, pr
- Location
- Paris, France
- Team size
- 501-1000
- Remote policy
- Hybrid
Overview
- Seniority
- Middle–Senior
- Remote policy
- Hybrid
- Location
- Paris, France
- Employment
- Full-time
- English
- C1
- Posted
- Aug 23, 2026