AI Data Trainer & Quality Analyst (Scale AI) — text data labeling and LLM quality evaluation
Trained and evaluated large language models by labeling thousands of daily text-related data points to improve downstream model performance for major tech clients. Performed sentiment analysis, entity recognition, and prompt response scoring to reduce overall error rates. Collaborated with global teams to refine labeling guidelines and onboard new remote annotators, while maintaining consistently high accuracy and quality scores. • Labeled 2,500+ data points daily • Achieved 99.2% accuracy • Reduced error rates by 35% • Ranked in top 5% and received Star Performer awards