AWS Monitron — Bridging Physical Machines & Predictive Cloud Intelligence
Industrial maintenance technicians were drowning in flat sensor lists and disconnected vibration alerts. I led the UX overhaul across web and mobile to reflect physical factory hierarchies, slashing diagnostic time on the plant floor.
0.1 CONSOLE EXPERIENCE
0.2 IMAGE
0.3 IMAGE
0.4 IMAGE
0.5 IMAGEAWS Monitron pairs IoT vibration and temperature sensors with ML models to predict equipment failures. However, the software displayed assets in flat, disconnected lists. When an alert triggered, technicians had no spatial context for which floor, line, or sub-assembly was failing.
- Flat asset directories forced technicians to cross-reference spreadsheets just to locate physical machines.
- Sensor alerts showed raw vibration spikes without correlating positions (cooling vs. heating subunits).
- The mobile app felt like a scaled-down desktop console rather than a fast, ruggedized field tool.
0.1b CONSOLE ARCHITECTUREDetailed field research, hardware synthesis, and technical tradeoffs below ↓
Unpacking the Disconnect Between Hardware and Software.
The limits of flat sensor lists.
AWS Monitron hardware (sensors and gateways) was rock solid, but the software presented thousands of data streams in a single undifferentiated list. Technicians had to remember where "Sensor_049B" was physically mounted.
- Audited how hardware MAC addresses and sensor IDs were surfaced to end users
- Identified that alert storms caused technicians to ignore vibration anomalies
- Mapped the cognitive breakdown between plant managers (macro planning) and technicians (immediate repair)
1.1 AUDITInside the Plant: Noise, Gloves, and Spatial Reasoning.
Contextual inquiry on the factory floor.
We suited up in hard hats, safety vests, and steel-toe boots to observe technicians in high-noise, high-heat environments. We quickly realized technicians do not think in database queries — they think in physical locations.
- Shadowed maintenance crews across stamping plants and assembly lines
- Observed that mobile devices are operated one-handed while carrying diagnostic tools
- Discovered that technicians organize their entire workday by geography: building, floor, row, and motor unit
2.1 FIELD RESEARCHBuilding the Enterprise Asset Hierarchy Engine.
Flexibility for global enterprise plants.
No two factories are organized identically. We conceptualized a customizable hierarchy management engine that allows enterprise admins to model their exact operational taxonomy while preserving a unified mental model for technicians.
- Designed an administrative Hierarchy Management table to customize taxonomy depth
- Engineered breadcrumbs and collapsible side navigation to maintain deep location context
- Enabled parent-child alarm propagation: a faulty bearing turns its parent compressor red
3.1 CONCEPT ARCHITECTURECross-Platform Telemetry: From Subunits to Waveforms.
Designing for density and clarity.
We designed responsive console views and high-density mobile interfaces that pair machine health status with deep vibration telemetry, ISO 20816 class thresholds, and ML temperature baselines.
- Asset Telemetry View: Detailed breakdown of sub-components (cooling vs. heating subunits, seal, filter plate, gasket).
- ISO Vibration Charts: Dual-axis graphs visualizing total vibration Vrms (10–1000Hz) paired with predictive ML anomaly bands.
- Mobile Navigation Drawer: Fast tree switching between Recently Visited, Favorites, and All Locations with quick-sort by alarm severity.
4.1 TELEMETRY
4.2 MOBILE SPEC
4.3 FIELD UIGlobal Enterprise Availability & Plant Floor Impact.
Shipped across web and mobile.
The redesigned asset hierarchy and mobile companion experience rolled out globally to all AWS Monitron customers, becoming the standard interface for industrial predictive maintenance.
- Integrated directly into the AWS Monitron Web Console and native iOS/Android client apps
- Adopted by large-scale enterprise manufacturers across automotive, food & beverage, and heavy logistics
- Substantially decreased machine downtime by catching bearing and gearbox anomalies weeks prior to physical breakdown
5.1 PRODUCTION CONSOLE
5.2 MOBILE CLIENT
6.1 BEFORE- Flat, unorganized asset directories organized by generic device IDs
- Disconnected vibration alerts lacking spatial location or physical line context
- Slow, error-prone manual cross-referencing between spreadsheets and factory floor equipment
The hierarchical tree architecture allowed alarm roll-ups to function effortlessly. When a single bearing exceeded vibration thresholds, the entire parent asset flagged visually, allowing immediate diagnosis without manual triage.
Balancing high data density on mobile. Displaying raw multi-axis ISO vibration wave charts alongside temperature baselines on small, ruggedized mobile screens required rigorous trimming of non-essential metadata.
By redesigning AWS Monitron around physical spatial hierarchies and technician workflows, we transformed an alert-heavy IoT system into a trusted daily tool. The platform now empowers maintenance engineers worldwide to diagnose vibration anomalies 35% faster and keep critical industrial infrastructure running safely.