What you'll do
Evaluate Arabic language model outputs and rank them by quality, accuracy, and safety (RLHF).
Write and review text in your native dialect — Gulf, Egyptian, Levantine, or Maghrebi.
Find vulnerabilities in Arabic models through adversarial testing and red-teaming.
Contribute to gold-standard Arabic datasets used to evaluate and improve AI models.
Requirements
Native Arabic speaker (any dialect), or fluent in English / other languages
Strong written communication skills
Attention to detail and consistency
Computer with reliable internet access
No prior annotation experience required
How it works
Create your account and tell us about your dialect, skills, and background.
Complete a short onboarding and qualification process to demonstrate your readiness.
Receive tasks, deliver quality work, and get paid for your contributions.