fatima zamil

I annotate and evaluate large-scale Arabic and English datasets for machine learning, focusing on semantic accuracy, grammar, and regional context. I also fact-check AI responses for healthcare, biology, and clinical scenarios against verified medical literature and protocols.

On multimodal projects, I quality-check and label images, audio logs, and video content—verifying bounding tags, temporal alignment, and audio transcriptions. I also grade and rank model outputs for RLHF-style priorities like accuracy, helpfulness, safety, and guideline adherence.

I’m careful about edge cases and ambiguous data, flagging logical discrepancies and providing detailed feedback to improve annotation guidelines. I review prompts across medical and policy-compliance domains to help align AI behavior with safety and helpfulness requirements.