AI Red Team – GenAI Adversarial Evaluation

The Adversarial AI Evaluation Scientist reviews and validates adversarial agent tasks, such as indirect prompt injection and data exfiltration, to ensure they are realistic, fair, and correctly graded. This role combines adversarial security judgment with statistical analysis and LLM-assisted tooling to identify quality issues, analyze attack patterns, and recommend improvements. The position requires strong analytical skills and expertise in ML security to evaluate model vulnerabilities and enhance AI safety.