Jev-IOT
Smart-meter fleet dashboard where code pre-filters telemetry and Jev classifies anomalies for remediation across millions of endpoints.
Overview
Jev IoT Telemetry Fleet Manager is a dashboard demo aimed at utilities and telecom providers managing millions of smart gas, water and electric meters. It uses a layered pipeline: a deterministic pre-filter, a Jev typed-question layer that classifies anomalies, and remediation logic, displayed on a 2.5D metro grid map with anomaly beacons and a feed of Jev decisions.
The README makes strong economic claims (for example monthly spend for 10 million meters and latency around 114 ms average) that are projections rather than measurements from a deployment. It runs locally with Node 18+, and npm test validates the pre-filter, the Jev contract and the remediation logic.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What kind of anomaly does this meter telemetry show?(inferred) | — |
| Yes / no | Does this anomaly need remediation?(inferred) | — |
How it works
- Ingest meter telemetry
- Deterministic pre-filter drops normal readings
- Jev classifies the remaining anomalies
- Apply remediation logic and show on the map
Strengths
- Pre-filter limits model calls
- Tests for each layer
- Visual operations dashboard
Caveats
- Cost and scale figures are projections
- Demo, not a deployed system
Getting started
`git clone https://github.com/pjmenon45/Jev-IOT.git && cd Jev-IOT && npm install && npm run dev`.At a glance
- Type
- Project
- Status
- prototype
- Language
- TypeScript
- Runtime
- Vite web dashboard
- Has code
- yes
- Has tests
- yes
Use cases and integrations
How curated lists describe it
Smart Utilities & Telecommunications: Ultra-low-cost, non-autoregressive AI telemetry classifier enabling sub-150ms anomaly triage and autonomic remediation across 10M+ smart meters for under $35/month.
yibie/awesome-jev
Frequently asked questions
What does Jev decide?
How to classify anomalous meter telemetry that passes the pre-filter.
Are the cost numbers measured?
They are projections in the README, not deployment measurements.
How do I run it?
npm install and npm run dev, then open localhost:5173.
Related projects
Home Assistant integration that runs weekly checks with Jev judgments and turns confident findings into Repairs cards for you to confirm.
ProjectbetaPythonScores OpenTelemetry logs with Jev for diagnostic value, priority and routing before any expensive LLM analysis.
ProjectbetaTypeScriptSplunk search command that asks Jev typed questions about events and adds probabilities, choices and scores as fields.
AppbetaPythonRust CLI and TUI that streams Kubernetes logs, applies offline rules and Jev analysis to flag what matters, then hands off.
ProjectbetaRustOpenTelemetry Collector processor that asks Jev how useful each metric is from its metadata and applies keep/drop policy.
ServiceprototypeGoOpenTelemetry trace processor that asks Jev whether an operation is diagnostic, business-critical and worth keeping, and annotates spans.
ProjectprototypeGo