← Back to projects
● Production▲ Vercel🚂 Railway
NT Battery RUL Monitoring System
1,944-battery LSTM MLOps platform for National Telecom UPS infrastructure
Production ML platform monitoring 1,944 UPS batteries across 9 NT data center facilities. Predicts Remaining Useful Life using LSTM neural networks (R²=0.85, MAE=2.3 days) and detects anomalies with Isolation Forest (F1=0.90). Full Medallion data pipeline, TimescaleDB hypertables, LINE Bot alerting with Gemini AI, and 3D facility visualization.
✓ Verified StackNotion: NT-POC Battery RUL Monitoring System v2.0, Jan 23 2026
TRY THE WORK
The system is deployed on edge nodes. Monitor container logs, memory allocations, and trigger virtual requests below:
edge-console://nt-battery-rul.dev
LIVE PREVIEW
Resource Monitor
EDGE NODE CPU12%
RAM ALLOCATION18.4MB / 128MB
PLATFORM CLUSTERVercel / Railway
KEY METRICS & FACTS
1,944
Batteries monitored
9
NT data center facilities
81
Battery strings
0.85
LSTM R² accuracy
2.3 days
MAE prediction error
0.90
Anomaly detection F1
10 seconds
Sensor ingestion interval
~18 seconds
Batch prediction (500 batteries)
50+ per timestep
LSTM features
243
Training experiments (GridSearchCV)
<150ms
API p95 latency
~80ms/battery
ML inference latency
DATA PIPELINE
Landing Zone → raw sensor ingest
→Bronze Zone → cleansed + deduped
→Silver Zone → enriched + normalized
→Gold Zone → ML features + analytics-ready