AI Data Annotation Specialist (Contract)
• Reviewed and labelled over 45,000 text samples across sentiment classification, named entity recognition (NER), and intent detection tasks for a large-scale conversational AI model, maintaining a personal quality score consistently above 97%. • Applied detailed annotation guidelines to evaluate model-generated responses for factual accuracy, tone appropriateness, and safety compliance under a Reinforcement Learning from Human Feedback (RLHF) framework. • Identified and escalated over 200 ambiguous edge-case samples to the QA lead, contributing to a guideline revision that improved inter-annotator agreement (IAA) by 11% across the annotation team. • Cross-validated labelled datasets against gold-standard benchmarks using Excel pivot tools, catching and correcting a systematic mislabelling pattern that affected approximately 3% of processed records. • Contributed to image-text alignment tasks, tagging 8,000+ image-caption pairs for visual question answering (VQA) model training, adhering strictly to bounding box and attribute labelling protocols.