What did I do?

• Discovery, research, design, prototypes, documentation.

Scope & Methodology

• Led the end-to-end product lifecycle, encompassing discovery, quantitative research, interaction design, and technical documentation.
• Collaborated with engineering teams to translate complex mathematical concepts into accessible user interfaces.

The Solution: Recursive Estimation

• Implemented Kalman filtering—a recursive algorithm utilised to estimate unknown variables—to mitigate signal noise and enhance the precision of body composition predictions.
• Transitioned the data model from single-point measurement to a multi-measurement aggregation, allowing for outlier detection and stochastic smoothing of results.

WITH KALMAN FILTERING

Filtered Estimate (Smooths out noise)

WITHOUT KALMAN FILTERING

Raw Measurements (No smoothing)

UX Challenges & Research

• Addressed the rigorous usability trade-offs required by the algorithm; the necessity for sequential triple-scans introduced significant friction and physical fatigue (Hick’s Law).
• Conducted time-on-task studies to quantify the impact of the extended workflow on user retention and satisfaction.
• Overhauled instructional design and on-screen guidance to scaffold the user through the intensified capture process and mitigate the increased barrier to entry.
• To help familiarise users with the concept, encourage regular scans, and communicate ‘the why’, the concept of the ‘calibration phase’ was introduced throughout copy.

Error Handling & Strategic Roadmap

• Identified critical failure states where cumulative error forced workflow restarts, prioritising these friction points for future technical mitigation.
• Defined the roadmap for a ‘Confidence Engine’ to parse environmental telemetry (lighting, pose, background complexity), reducing reliance on perfect user execution and addressing ‘Garbage In, Garbage Out’ (GIGO) constraints.