Data Annotator (Contract/Remote)
As a Data Annotator at iMerit Technology, I performed Reinforcement Learning from Human Feedback (RLHF) by ranking multimodal AI outputs, including text, image, and audio. My role included prompt classification and rating for truthfulness, reasoning, and safety to reduce model hallucinations. I contributed by conducting reverse prompt engineering, creating ideal response baselines, and writing video-to-text and image generation prompts. • Evaluated generative AI outputs across multiple data types every day. • Used comparative judgment techniques to assess AI coherence and compliance with guidelines. • Ensured all outputs were copyright-free and aligned with project rules. • Applied strict adherence to annotation standards and quality metrics.