Data Annotation Specialist – RWS
Executed high-volume labeling projects to generate and assess multimodal labels across image, text, and audio. Conducted multi-stage audits to identify and correct labeling bias and inconsistencies while iteratively improving model outputs.• Applied complex classification schemas across image, text, and audio datasets.• Ranked and refined model outputs to improve factualness, safety, and helpfulness.• Investigated model failures and edge cases to guide iterative training cycles.• Evaluated model performance using standard metrics and updated moderation guidance for misinformation and hate speech.