From data to decisions. From the field to value.

An integrated IoT platform that collects Wi-Fi CSI, temperature and humidity and analyzes unusual signal patterns. Customers, sites, locations and devices are organized with access permissions.

• ESP32 · MQTT sensor collection
• V2 activity/night model training and versioning
• Feature generation and inference about every 5 minutes
• Map monitoring · collection, training and job status

Anomaly scores describe environmental signal changes. They do not establish movement, falls, sleep states or medical risk.

AI visual inspection starting with a USB camera and Windows PC. Built with Python, PySide6 and YOLO11, it trains on product images and connects OK/NG/hold decisions with inspection records.

• Capture/import images · confirm product/LOT
• Product-specific YOLO11 image classifiers
• Resident AI worker · model cache · local inference
• SQLite inspection records · product statistics dashboard

The current model classifies images. Defect localization, mobile inspection and MES/production-line integration are future extensions that require field validation.