Personal Statement
Motivation
When our recommendation model failed for cold-start users, I rebuilt the evaluation set and traced the gap to sparse behavioural signals. During my AI product internship, I learned that responsible deployment requires more than strong offline metrics.
Evidence
My perspective is grounded in an AI product internship, multimodal recommendation research, a machine learning project, and an exchange semester at the National University of Singapore.
Programme fit
HKU's MSc in Artificial Intelligence would help me connect rigorous model evaluation with the design of intelligent systems that people can understand and trust.
Direction
I want to build AI products whose value can be measured not only by model performance, but also by the quality of the decisions they enable.







