Independent Multimodal AI Evaluation — Remote Freelance
Evaluated AI-generated text, audio, and multimedia outputs using rubric-based scoring frameworks. Compared outputs to select the most coherent, contextually accurate, and natural responses. Detected hallucinations, logical inconsistencies, fluency issues, and weak contextual understanding to improve training data quality. • Performed high-volume rubric-based evaluations • Compared multiple AI responses for coherence and context • Flagged hallucinations and logical or fluency problems • Assessed naturalness and contextual understanding across modalities