ilyas sayh

I’m a detail-oriented Data Annotator and AI Evaluator with hands-on experience in data labeling, AI response evaluation, and quality assurance across text, image, and structured data projects. I follow annotation guidelines closely, review outputs for accuracy and consistency, and correct errors before submission to keep data reliable.

In RLHF-focused work, I complete ranking and pairwise comparison tasks across [French, Arabic, English], identifying factual errors, hallucinations, and guideline violations. I run quality assurance and validation checks before submission so preferences and outputs align with the intended behavior.

I also evaluate chatbot responses for helpfulness, correctness, and safety using structured review criteria, writing clear justifications for evaluation decisions. I flag biased, toxic, and unsafe content, focusing on safety-focused evaluation and content moderation consistency checks, so teams can improve response reliability. I’m especially interested in roles centered on annotation quality, guideline adherence, and AI data review.