Remote AI Safety Red Teamer - Adversarial Prompting

Mercor is seeking an AI Safety Red Teamer for contract work to stress-test frontier AI models and identify unsafe behaviors. The role emphasizes documenting vulnerabilities and contributing to red-teaming efforts for safety benchmarking.

You will design adversarial prompts, evaluate model robustness, and collaborate with AI researchers to improve alignment across diverse, sensitive domains. Remote contract with competitive hourly compensation.