Negotiable
Undetermined
Remote
Remote
Summary: The Red Team Tester will be responsible for conducting testing of AI models to identify vulnerabilities and develop evaluation frameworks. This role involves collaborating with global teams and member firms to communicate risks and improve security posture against emerging AI threats. The position requires a strong background in penetration testing and offensive cyber operations, particularly in the context of AI. The tester will also create adversarial datasets and benchmarks to evaluate model robustness under various attack conditions.
Key Responsibilities:
- Design and execute comprehensive adversarial testing campaigns against AI models.
- Research attack vectors and prompt injection techniques to identify model vulnerabilities.
- Conduct red team exercises simulating real-world deployment scenarios and edge cases.
- Create adversarial datasets and benchmarks to evaluate model robustness.
- Collaborate with security teams to perform red teaming activities and document findings.
- Develop and maintain process documentation and reports on testing activities.
Key Skills:
- Minimum 5 years of penetration testing, red team operations, or AI Red Team experience.
- Degree in computer science, information technology, programming, development, Artificial Intelligence, or cybersecurity.
- Background in AI red teaming or offensive cyber operations.
- Understanding of Artificial Intelligence, Machine Learning, software applications, cloud computing, and networking.
- Excellent communication and stakeholder management skills.
Salary (Rate): undetermined
City: undetermined
Country: undetermined
Working Arrangements: remote
IR35 Status: undetermined
Seniority Level: undetermined
Industry: IT
The Red Team Tester will conduct testing of AI models to find vulnerabilities, develop new evaluation frameworks, communicate risks to stakeholders across the global teams and the network of member firms, and collaborate with defensive and development teams.
Key accountabilities % of time Accountability 10 Design and execute comprehensive adversarial testing campaigns against AI models, including but not limited to large language models, multimodal systems, and autonomous agents. 10 Research attack vectors and prompt injection techniques to identify model vulnerabilities, jailbreaks, and unintended behaviors. Evaluate and introduce AI Security tools that decrease mean time to detect and respond to AI-specific threats.
50 Conduct red team exercises simulating real-world deployment scenarios and edge cases. Systematically probe for bias, toxicity, misinformation generation, and other harmful outputs across diverse contexts and demographics. Improve firm s security posture against emerging AI threats. Concentration is on executing standardized security tests.
10 Create adversarial datasets and benchmarks to evaluate model robustness under various attack conditions.
10 Collaborate with security teams (e.g., GSOC, member firm security teams, developers) to perform red teaming activities, and document and present findings, vulnerabilities, and remediation recommendations to drive the mitigation of identified risks. 10 Develop and maintain process documentation, create reports on testing activities, track operational metrics, and perform other programmatic tasks as required to support the AI Red Team's function. 100 Total FTE (100% maximum)
This is the job for you if you have [experience / knowledge / qualifications] Minimum 5 years of penetration testing, red team operations, or AI Red Team experience.
Computer science, information technology, programming, development, Artificial Intelligence, or cybersecurity degree (Bachelor's or higher preferred) from an accredited college or university or equivalent work experience.
Background in AI red teaming, red teaming, or offensive cyber operations. Understanding of Artificial Intelligence, Machine Learning, software applications, cloud computing, and networking. Excellent communication and stakeholder management skills. Ability to work effectively across geographies and t