Expert Talent Pool, Safety Testing (Expert)

Responsibilities

• Design and lead red teaming campaigns for high-risk model capabilities
• Develop and document jailbreak techniques for reproducibility
• Test agent systems for prompt injection and tool misuse
• Design bias testing frameworks across demographic groups and languages
• Audit model behavior against written safety policies
• Write moderation guidelines and review other moderators' work
• Rate findings by severity and recommend fixes

Requirements

• At least five years in security testing, trust and safety, policy, or fairness research
• Hands-on AI safety testing experience
• Experience designing test plans, threat models or audits
• Experience leading or reviewing testers or moderators
• Familiarity with Burp Suite, Garak, PyRIT or OWASP LLM Top 10
• Strong written English and ability to handle sensitive content

Benefits

• Project-based pay of $75 to $175 per hour, paid for hours worked
• Fully remote, with projects from one week to several months