17 Nov
|
Cohere
|
Moncton
Use Trust & Safety judgment to evaluate model outputs for Cohere, working remotely as a part-time independent contractor in Canada or the United States. Written English is essential.This Data Annotation Specialist role focuses on safety evaluation for Large Language Models supporting AI systems. You will recognize risks across user intents, apply detailed rubrics, and explain why outputs are protected, unsafe, high-quality, or low-quality. The work includes sensitive material and requires cultural awareness, bias sensitivity, and emotional resilience.Key Responsibilities:Assess model outputs against safety guidelines and style standardsEvaluate content across sensitive, ambiguous,
and context-dependent scenariosWrite realistic prompts and adversarial examples for safety testingAudit annotations and flag unclear or inconsistent guidelinesDeliver consistent decisions across high-volume evaluation tasksRequirements:One or more years in Trust & Safety, moderation, or related analysisStrong contextual and sociocultural judgment with bias awarenessEmotional resilience with explicit, violent, or psychologically disturbing contentExcellent written English and ability to justify evaluations clearlyComfort working independently with new tools remotelyBring nuanced safety analysis, cultural judgment, and clear writing to Cohere's AI evaluation and data quality work.
📌 Trust And Safety Ai Evaluator (Moncton)
🏢 Cohere
📍 Moncton