Data Annotation Specialist, Arabic Language Najdi/Hijazi Dialect

Stack

Responsibilities

• ▹Label and rank machine learning data with advanced Arabic (Najdi/Hijazi) proficiency, ensuring data integrity and quality
• ▹Audit and correct inaccuracies in language data (text, image, JSON, TSV/CSV, Markdown)
• ▹Efficiently complete reading and text-based assignments with high attention to detail
• ▹Evaluate and complete preference-based tasks, assessing which Arabic (Najdi/Hijazi) responses best conform to project guidelines
• ▹Identify and recommend opportunities to improve data quality and model performance in Arabic-language tasks
• ▹Collaborate and provide feedback to cross-functional team members on Arabic-language matters
• ▹Note: the role may involve exposure to NSFW (not-safe-for-work) text content as part of the annotation work

Requirements

• ▹Native or near-native proficiency in Arabic (Najdi/Hijazi); excellent command of English
• ▹Expert reading and writing skills, provable via a written test
• ▹Basic familiarity with JSON/TSV/CSV/Markdown data manipulation
• ▹Strong attention to detail and commitment to accuracy
• ▹High tolerance for repetitive and monotonous work
• ▹Ability to follow complex instructions, navigate ambiguity and work independently
• ▹Strong sense of urgency and time management
• ▹A reliable laptop (BYOD)
• ▹Available to commit at least 16 hours per week; based in Canada

Nice to have

• ▹Curiosity about technology and a knack for creative problem-solving

Soft skills

What we offer

• ▹30 CAD/hour contract rate
• ▹Performance incentives
• ▹Fully remote work within Canada
• ▹Bring-your-own-device (BYOD) flexibility
• ▹12-month independent contractor engagement

About the company

Cohere's mission is to scale intelligence to serve humanity, training and deploying frontier models for developers and enterprises building AI systems such as content generation, semantic search, RAG and agents.