GenAI CBRNE Cyber Expert

ActiveFence is seeking a GenAI CBRNE Cyber Expert to lead adversarial red-teaming of generative AI systems against high-risk cyber threats involving chemical, biological, radiological, nuclear, and explosive (CBRNE) domains. This remote role involves designing sophisticated jailbreaks, prompt mutations, and multi-turn attack scenarios to identify where GenAI systems could inadvertently enable cyber-CBRNE threats. You will work with AI safety and model-alignment teams to classify vulnerabilities, develop test suites, and provide actionable recommendations.

• Execute rigorous adversarial red-teaming of GenAI models across cyber-CBRNE threat scenarios.
• Design and execute adversarial prompts, jailbreaks, prompt mutations, multi-turn conversations, and scenario-based evaluations.
• Develop realistic cyber-CBRNE attack scenarios involving critical infrastructure, ICS/OT, and hazardous-material facilities.
• Evaluate whether models can be manipulated into assisting threat actors through attack planning, vulnerability analysis, and aggregation of benign information.
• Identify and document failure modes including indirect requests, role-playing, encoded prompts, and multi-turn escalation.
• Conduct systematic taxonomy audits, classify failures by severity, and provide recommendations to AI safety teams.
• Develop repeatable red-team test suites, adversarial datasets, evaluation rubrics, and risk taxonomies.