After 20 years optimizing 5G and 4G networks for T-Mobile, Verizon, AT&T, Bell Canada, Nokia, and Mavenir - I decided to build my own AI-powered NOC tool from scratch using real network data.
Upload any Nokia, Ericsson, or Samsung KPI export. Get every cell ranked by health score in seconds. Click any cell - see exactly which KPIs are flagged
and why, with 14 days of trend history.Ask the AI assistant what’s wrong and what to do.
What it found on a real 5G network:→ 2,036 unique cells analyzed across 14 periods
→ 8 critical, 74 degraded, 123 anomalous
→ Cell_0319: SINR flagged, HO Success Rate 0%,
Abnormal Release trending from 11.7% to 18.57% over 14 days - field investigation dispatched
Three ML models under the hood - Isolation Forest anomaly detection, Random Forest handover failure classification, and coverage prediction on 567,195 real LTE drive test measurements.
Deployed with FastAPI, a GPT-4o network assistant that answers questions about your live network data, and a one-click Windows launcher.
Open to AI-RAN, Network Operations AI, and Telecom
ML Engineering roles.
LinkedIn: 
