01 — Rednote AI Trust
Led research and design on trust loss in Xiaohongshu's Ask AI feature.
Proposed a dual framework of explainability + user control. Delivered four-phase interaction design, validated through user testing.
02 — One-shot Prompting
Designed and ran a controlled A/B experiment to quantify how one-shot prompt strategies affect LLM output action-orientation.
Statistically significant result (U = 0.00, p < .001, r = 1.00).
