Human–AI Interaction in Product Design

My Berkeley MIMS capstone investigates a question every product team is now living: when creative intent passes through AI “vibecoding” tools, where does the translation break — and how should development workflows change to keep human taste in the loop?

• Studied product teams across Silicon Valley corporations to map where intent-to-output fidelity degrades.
• Applied mixed-methods research to surface concrete failure points in AI-assisted product development.
• Translated findings into recommendations for improving AI-assisted design and engineering workflows.