Analyste Red Teaming des systèmes multi-agents - Risques - La Defense, Ile de France, France
Core
Conduct literature reviews, develop metrics and dashboards, and execute red teaming campaigns to identify vulnerabilities and failure scenarios in Large Language Model (LLM) based multi-agent systems.
Role type
Junior IC model risk analyst (multi-agent systems)
Builds
Monitoring frameworks, evaluation metrics, and dashboards for agent system performance and stability
Domain
Banking / AI Safety / Multi-Agent Systems
Deliverable
production ML models
Required skills
Machine Learning, Deep Learning, Python, Red Teaming, System Monitoring, Metrics Development, Documentation
Preferred skills
Literature Review, Presentation Skills, Autonomous Work, Team Collaboration
Technologies
Python, Open Source Agent Frameworks, Large Language Models (LLM)
Responsibilities
Conduct literature reviews on monitoring and red teaming approaches for multi-agent systems; Develop metrics and dashboards to track agent interaction performance and stability; Design and execute red teaming campaigns to identify vulnerabilities and reasoning errors; Compare monitoring strategies to identify best practices; Document findings and present results to the team.
Seniority
Junior, internship
