Overview
ARGVS AI
ARGVS AI is a multi-modal AI security platform I built from scratch — a single system that understands live camera and audio feeds and unifies security functions that are normally sold as separate, expensive products. One intelligent core scales from a single camera to a facility with hundreds, keeping all data on-site.
01 / The Idea
Security today is fragmented and expensive — license plates, faces, fire, and workplace safety are usually separate systems, most of them cloud-dependent. ARGVS AI replaces that scattered setup with one local-first platform: all analysis runs on the device, no cloud required, so it stays fast, affordable, and privacy-respecting.
02 / What It Does
03 / Real-World Scenarios
A worker enters a hazard zone without a helmet or vest.
PPE detection catches it instantly — the event is logged and an alert is sent.
A lone worker falls or collapses on site.
Pose analysis detects the fall and raises an immediate alarm.
An unauthorized person enters the site at night.
Face recognition flags them as "unknown" and sends a photo alert.
A small fire or smoke starts in an empty warehouse at night.
The fire/smoke classifier detects it early and alarms before it spreads.
An unregistered vehicle approaches an entrance.
The plate is read and checked — unauthorized vehicles are flagged and logged.
A camera is covered or sabotaged.
Tamper detection notices the blackout and alerts immediately.
04 / Design Principles
- Local-First
Data stays on-site — privacy-respecting by design.
- One Platform
Many capabilities in a single, unified system.
- Edge-Ready
Scales from a single PC to embedded devices.
- Real-Time
Sees and understands the environment live.
05 / Why It Matters
ARGVS AI sits at the intersection of my two disciplines — the spatial thinking of architecture and the systems thinking of building AI that runs reliably in the real world. It's designed to work across very different contexts: construction sites, residential complexes, factories, parking areas, and homes.