Arabic Document Specialists – Data Sourcing for AI Training

We are seeking detail-oriented freelancers to contribute to an AI training data sourcing project aimed at developing advanced multilingual AI models. In this role, you will work with document templates in the Arabic language. The data collection involves three categories: Forms, Non-Forms, and Digital formats.

📌 Project Details:

Forms:

One base template will be provided and must be used to generate eight unique variations. Each variation must preserve the structural integrity and formatting standards outlined in the project guidelines.

Non-Forms:

Two templates will be used, and each submission must achieve a minimum of 90% key-value alignment.

Digital Forms:

These follow the same rules and formatting standards as Forms, with an emphasis on structured digital content representation.

⚠️ Important Requirements:

✅ Document Length: Only single-page and two-page template-based documents are accepted in any category (Forms, Non-Forms, or Digital).

✅ Data Quality: Submissions must include 100% realistic, well-formatted, and clean data. Use of placeholder text (e.g., “lorem ipsum”) is strictly prohibited.

✅ Language Accuracy: All submissions must demonstrate 100% Arabic language quality, appropriate to the document’s structure and domain.

✅ Quality Assurance: If a file is flagged for quality issues, it is your responsibility to revise and resubmit it. Multiple quality issues may affect your eligibility for upcoming work.

✅ Domain Coverage: The project spans nearly 100 unique domains, such as legal, medical, financial, education, and more — with additional domains expected to be added in the future. Familiarity with various sectors will be beneficial.