Data Annotator / AI Trainer (RLHF) at Luel AI
Provided large-scale dataset annotation for text, conversational exchanges, and structured data to support AI training and fine-tuning. Conducted RLHF response evaluation by rating AI outputs for accuracy, relevance, coherence, and safety. Maintained high annotation quality by consistently achieving accuracy above 95% across multiple concurrent tasks. • Labeled and organized text and dialogue data for model training pipelines. • Evaluated response quality as part of RLHF workflows. • Refined prompts to improve task alignment and response quality. • Updated annotation guidelines with cross-functional collaborators to improve consistency.