Freelance Graphic/Audio Professional & AI Data Evaluator (RLHF workflows and AI output evaluation)
Provided RLHF-style evaluations for AI-generated content by applying pairwise ranking and rubric-based scoring to align outputs with human aesthetic, factual, and creative standards. Performed qualitative assessment of AI-generated audio, visual, and text content to refine model accuracy and improve prompt adherence. Contributed to multimodal verification to ensure generated material remained contextually sound and aligned with professional design expectations. • Conducted pairwise ranking of candidate outputs against human-guided rubrics • Scored items using rubric-based criteria for both creative and factual quality • Reviewed multimodal generations (audio/visual/text) for context and alignment • Supplied qualitative feedback to guide iterative model improvement