Safety Data Annotation Specialist

Apply safety judgment and written English skills remotely as a part-time Data Annotation Specialist with Cohere in Canada or the United States.

This independent contractor role supports large language model safety through annotation, evaluation, and quality improvement. You will assess sensitive outputs, apply detailed guidelines, create test cases, and explain decisions consistently. The work requires experience in content moderation, trust and safety, AI data annotation, or LLM evaluation.

Key Responsibilities:

• Label, rank, audit, and refine generated text
• Apply safety guidelines to ambiguous model outputs
• Create prompts, scenarios, and adversarial safety tests
• Identify inconsistencies and provide calibration feedback
• Complete precise tasks independently in asynchronous workflows

Requirements:

• At least one year of relevant analytical experience
• Strong contextual, sociocultural, and bias awareness
• Comfort evaluating explicit, toxic, or disturbing content
• Excellent written English and evaluation explanations
• Accuracy across high-volume, repetitive annotation tasks

Bring careful safety evaluation, content judgment, and independent execution to Cohere’s model improvement work.