This entry-level contractor role involves reviewing language-model responses in social scenarios, applying evaluation rubrics, and providing structured feedback in English to improve AI system behavior. The position is open worldwide, requires at least 20 hours per week for a 2- to 3-week project, and tasks take approximately 60 minutes each. Candidates must be fluent in English and possess strong written communication skills and attention to detail.