Ahmed Salah

I evaluate Arabic AI-generated responses at Outlier AI, comparing multiple model outputs against detailed evaluation guidelines.

I assess instruction following, reasoning quality, language accuracy, and overall response quality. I provide structured, objective feedback that helps improve model performance.

I also identify inconsistencies, edge cases, and quality issues while working independently in a remote environment and meeting project standards.