AI & Scientific Data Support — Independent Work (AI evaluation and biomedical text annotation)
Evaluated AI-generated scientific content for accuracy, coherence, and consistency with biomedical evidence while identifying errors and hallucinations. Annotated biomedical text by classifying scientific concepts using domain-specific labeling guidelines and prompt-based test cases. Provided structured feedback to improve AI reasoning in scientific contexts. • Labeled and classified biomedical concepts from text • Performed factuality and reasoning error checks on AI outputs • Flagged hallucinations and inconsistencies for correction • Delivered actionable improvement notes for AI performance