AI Jailbreak & Prompt-Injection Security Expert - Remote

Location:

United States

Role Type: Contractor
Location: Remote

We are seeking experienced AI Security Experts to evaluate AI systems for jailbreak attacks, prompt injection, and security vulnerabilities. The role focuses on AI safety testing, red teaming, and developing frameworks to improve model security and reliability.

• Conduct AI safety evaluations and LLM red teaming.
• Identify jailbreak vulnerabilities and prompt-injection risks.
• Develop adversarial testing strategies and evaluation frameworks.
• Create and maintain security regression test suites.
• Assess AI tool-use abuse and potential security threats.
• Document findings and recommend improvements to AI safety.
• Collaborate with technical teams on risk mitigation.

• Minimum 2 years of experience in AI security, adversarial machine learning, or LLM red teaming.
• Knowledge of jailbreak techniques, prompt injection, and adversarial AI attacks.
• Advanced degree in Computer Science, Cybersecurity, Machine Learning, or a related field, or equivalent experience.
• Familiarity with LLM architectures, prompt engineering, and security assessment tools.
• Strong analytical, research, and communication skills.
• Published research, open-source security tools, or conference presentations are preferred.