AI Safety & Red Team Evaluator — Short-Term Sprint

About the role

We're assembling a focused cohort of domain experts to support a fast-paced AI safety evaluation sprint. This is a short-term, high-intensity engagement designed for professionals who can quickly assess model behavior, identify risks, and apply structured judgment under time pressure. The work centers on stress-testing AI systems for safety, robustness, and policy alignment through hands-on red teaming and adversarial evaluation.

This is an ideal opportunity for specialists in trust & safety, cybersecurity, content moderation, or AI policy who want meaningful, well-compensated project work without a long-term commitment.

What you'll do

• Conduct structured red-teaming and adversarial testing sessions against AI model outputs
• Evaluate responses for safety violations, policy breaches, and harmful edge cases
• Document findings with precision, using strong written English and clear reasoning
• Apply content moderation and trust & safety judgment to ambiguous or adversarial scenarios
• Collaborate closely with a coordination team during a compressed, fast-moving sprint window
• Maintain reliable, responsive availability across consecutive working days to keep pace with the sprint schedule

Requirements

• Demonstrated experience in AI safety, red teaming, cybersecurity, trust & safety, content moderation, policy evaluation, or adversarial testing
• Strong written English communication and meticulous attention to detail
• Ability to work independently while staying responsive and coordinated with the broader team
• Comfort operating in a fast-paced, deadline-driven sprint environment
• Reliable internet connection and availability for concentrated blocks of time over consecutive days

Compensation

This engagement pays a flat hourly rate of **$45/hour**, with total hours determined by sprint duration and scope. Payment is processed through the platform upon verified completion of assigned sessions.