LLM Safety Evaluator (Hebrew & English Required)

In this fully remote, hourly contractor role, you will review AI-generated responses and generate safety-focused evaluation content in English and Hebrew. Responsibilities include curating and labeling adversarial or safety-sensitive training examples, reviewing and scoring model outputs, documenting safety failures, and stress-testing models for policy gaps. You will help ensure outputs are accurate, safe, and clearly explained to prevent generation of unintentional or unsafe content, with tasks sometimes involving explicit or sensitive material. Your feedback and evaluations will directly impact the training and safety of large language models for a global AI data services organization.