AI Trainer, Uber AI (AI Trainer)
Created and curated training datasets used for LLM fine-tuning and RLHF pipelines. Evaluated and ranked model outputs to measure alignment, accuracy, and safety improvements. Annotated multiple text-based formats including structured content and conversational samples to support supervised learning. • Built training data for LLM fine-tuning/RLHF pipelines • Performed evaluation and output ranking for quality and safety • Labeled/annotated conversational and structured text examples • Developed annotation guidelines and QA standards with cross-functional teams