Research Engineer - Cybersecurity AI for RL Environments

Mistral; Paris; Sur place; EUR 70 000 - 105 000; Plein temps

Résumé du poste

Mistral in Paris is seeking a Research Engineer to advance cybersecurity capabilities inside our AI systems. You will design RL environments that simulate realistic security scenarios and build scalable pipelines to generate synthetic training data for model improvement.

You will own infrastructure behind environments—containers, VMs, and Kubernetes—while collaborating with researchers and security engineers. Proficiency in Python and a pragmatic research-to-engineering mindset are essential.

Qualifications

• Hands-on experience in offensive and/or defensive cybersecurity.
• Strong software engineering skills with clean, well-tested code.
• Fluency in Python; comfortable reading C/C++ for vulnerability analysis.
• DevOps experience: Docker, Kubernetes, cloud deployments.
• Pragmatic research-to-engineering mindset: ship first version, then scale.
• Familiarity with RL techniques and LLM training methodologies is a plus.

Responsabilités

• Design and implement RL environments simulating cybersecurity scenarios.
• Build pipelines and infrastructure that generate synthetic cybersecurity training data at scale.
• Conduct experiments and evaluations of model capabilities in these environments.
• Own the infrastructure behind the environments: containers, sandboxes, VMs, Kubernetes, as code.
• Collaborate with researchers and security specialists to translate domain knowledge into reproducible environments.
• Write clear, efficient Python code with strong testing, review, and CI/CD practices.

Connaissances

Offensive/defensive cybersecurity

Software engineering

Python

RL/LLM training

Self-starter & collaborative

C/C++ (reading)

DevOps

Outils

Docker

Kubernetes

Description du poste

Mistral in Paris is seeking a Research Engineer to advance cybersecurity capabilities inside our AI systems. You will design RL environments that simulate realistic security scenarios and build scalable pipelines to generate synthetic training data for model improvement.

You will own infrastructure behind environments—containers, VMs, and Kubernetes—while collaborating with researchers and security engineers. Proficiency in Python and a pragmatic research-to-engineering mindset are essential.