VENTURUS.

Venturus builds AI-powered solutions for a major international manufacturing client. Embedded in a cross-functional product team for 4 years, I led UX for an AI quality control tool used across the factory floor — from operators reviewing flagged defects to managers overseeing full production runs.

• 01Mapped the review workflow end-to-end — traced where the two months were actually going: flag, switch screens for context, decide, repeat.
• 02Reframed the problem — from "review every image" to "find what's actually abnormal." The bottleneck was prioritization, not decision speed.
• 03Designed a histogram view — operators see the full batch distribution at a glance, filtering and revealing patterns simultaneously.
• 04Eliminated multi-tool switching — consolidated context directly into the review interface, removing a structural source of delay.
• 05Advocated for design system adoption — extended the same visual logic across modules once the histogram flow was validated.

A redesigned quality control interface centered on a histogram view that surfaces statistical outliers instead of presenting every image as equally worth reviewing. Operators triage by pattern first, drilling into specific defects only when the data calls for it. Key screens: production run histogram overview, outlier drill-down view, defect confirmation flow.

Full production run review time dropped from roughly two months to two to three weeks — close to an 85% reduction in review cycle time. The shift came from changing what operators were looking at, not how fast they could click through it: replacing a flat list with a statistical view turned review from exhaustive to targeted.