A desk lamp that detects poor posture in real time and uses ambient light + motion to nudge users to self-correct—no wearables, no pop-ups.

Role: Product / UX Designer (Interaction, Visual Design, Research, Software Development)

Team: Shrey Patel, Keira Wong, Louis Santos

Timeline: 6 weeks

Tools: Figma, p5.js, ml5.js (PoseNet), Arduino, FAST_LED, After Effects

My Contribution

• Designed the interaction model mapping posture states to feedback states
• Built the p5.js + PoseNet visualization that mirrors posture quality on screen
• Led visual and experience decisions to keep feedback clear and non-intrusive
• Supported testing and iteration based on user behavior and system reliability

Problem

Remote workers and students spend long hours seated and often slip into poor posture. Existing posture solutions rely on wearables or notifications, which are intrusive, easy to ignore, and hard to sustain long-term.

Goal

Design a posture-awareness product that:

• Detects posture continuously and in real time
• Provides feedback that is subtle, immediate, and non-disruptive
• Fits naturally into an existing desk setup with minimal adoption friction