DeepInspect Cloud

• Users monitor, root-cause and improve production performance here.
• Built to monitor the performance of multiple vision systems at once.
• Tracks OEE% and analytics for each plant.
• Surfaces rejection-ratio trends and inspection images to root-cause production and machine issues.

Make it easy for end users to monitor plant performance — and to understand why it changed.

I led the design of DeepInspect Cloud from scratch, simplifying genuinely complex manufacturing data into a dashboard that a plant head and an operator could both read at a glance.

The hardest part was making real-time vision-system health, OEE% and rejection trends feel trustworthy rather than overwhelming.

Same Double Diamond and Lean UX process as Train, with a sharper focus on data hierarchy, multi-stakeholder readability and continuous iteration into hi-fi design and development.

Before the design team existed, users had no digital platform at all for monitoring productivity or manufacturing performance. I interviewed our primary users to understand life without a monitoring application.

• No visibility into real-time performance — no live metrics, so inefficiencies on the shop floor went undetected
• Manual data checks for OEE% — operators pulled spreadsheets and logs by hand, slowly and error-prone
• Delayed root cause analysis — diagnosing drops meant digging through fragmented data across systems and departments
• Inconsistent productivity tracking — every team used different tools and formats, so reporting was unreliable
• Reactive decision-making — problems were found late, so decisions cost time and resources
• No accountability or traceability — no audit trail to learn from past issues

• Interviews with 4 users plus competitor analysis of Elementary ML, Keyence, Cognex and DesignX
• 82% of users felt a monitoring application would make their work substantially easier
• 36% were frustrated that data mismatched when they calculated each vision system's productivity manually

• A centralised monitoring dashboard with real-time visibility per vision system and aggregated plant-level performance — no manual tracking
• An intuitive root-cause analysis flow for spotting performance drops and defect patterns across machines and lines
• Clear, visual OEE% insights so teams can compare effectiveness without spreadsheets
• A scalable information architecture that moves seamlessly from a single asset to plant-wide trends
• Filters, drill-downs and comparison tools for proactive advanced analytics
• Competitor benchmarking to fill experience gaps existing solutions ignored
• Prototypes tested across roles, from operators to plant heads
• A modular UI system that supports future scale while staying consistent

High-fidelity mockups in Figma, specced closely with the front-end team for interactions the mockups didn't cover, and a UI/UX review of every front-end ticket before release.

Factory teams gained usable, trustworthy data visibility — and feedback from users and sales highlighted how much faster their end-to-end monitoring and root-causing became.

• Complex data needs ruthless hierarchy. The value is in what you choose not to show first.
• Design for the whole ladder of users — an operator and a plant head need the same truth at different resolutions.