OCCDEC AI Matrix

OCCDEC AI Matrix gives you a real-time window into your AI models. Monitor accuracy, training loss, prediction confidence, and performance metrics from anywhere on your phone.

Key Features:

Performance Dashboard — View all your key AI metrics at a glance. Track accuracy, latency, prediction volume, and uptime with live-updating KPI cards and trend charts.

AI Matrix Visualization — Interactive heatmap showing cross-model performance correlations. Instantly spot underperforming models, anomalies, and optimization opportunities across your entire AI infrastructure.

Model Statistics — Deep-dive into individual model performance with precision, recall, F1 scores, and confusion matrices. Track training run history with success/failure status and duration.

Training Monitor — Follow active training runs in real-time. Get notified when training completes, fails, or when performance metrics cross defined thresholds.

Multi-Model Support — Monitor multiple AI models simultaneously including object detection, face recognition, anomaly detection, and custom classifiers. Compare performance side-by-side.

Customizable Alerts — Set up push notifications for performance degradation, anomaly spikes, training completion, and model health events. Never miss a critical AI infrastructure event.

Offline Support — Data is cached locally so you can review recent metrics even without an internet connection. Automatic sync when connectivity is restored.

Built by OCCDEC — Enterprise-grade AI monitoring powered by cloud infrastructure. Part of the OCCDEC platform developed by LineSpotting AB in Sweden.

Requirements:
- Android 6.0 or later
- Internet connection for live data
- No login required — uses API key authentication

Privacy: AI Matrix is a read-only monitoring dashboard. No personal data is collected, stored, or transmitted. The app only reads aggregated model performance metrics from the OCCDEC cloud API.

Recent changes:
Live metrics fixes, sample-data fallback so all dashboards show content, refreshed dark design, and stability improvements.