My research examines the limits of perceptual classification in beauty-related AI by integrating empirical modeling, systematic review, and human-agreement measurement. Through this work, I aim to develop evaluation frameworks that account for perceptual variability rather than assuming fixed ground truth. In the future, I plan to pursue a research-oriented career in industry, particularly in AI product development or applied research. I hope to contribute to the design of responsible, evidence-based AI systems that engage with socially constructed concepts such as beauty and identity.