민수 지

I'm evaluating and ranking LLM responses at Outlier AI for accuracy, coherence, safety, and guideline adherence, while writing nuanced prompts and gold-standard reference answers for RLHF datasets.

I've supported AI training at RWS through data annotation and quality evaluation, with a focus on accurate, consistent labeling. I identify hallucinations, biases, and structural errors and provide analytical justifications that support model alignment and fine-tuning.

My background also includes developing and reviewing English learning materials at Edubest and Avalon Academy, designing community programs at Jungnang-gu Gender Equality Center, and coordinating cultural content projects. As a native Korean speaker fluent in English, I bring cross-cultural communication, content evaluation, and careful documentation to AI data work.