Manufacturing
Safety & Productivity
Vision AI that enforces compliance and surfaces lost productivity without adding headcount
Manual monitoring couldn't keep up with safety compliance on an active production floor, PPE violations went unnoticed, idle equipment quietly eroded productivity. I built a GPU-accelerated computer vision system wired into their existing camera infrastructure. It detects safety violations in real time, flags idle equipment, and turns hours of footage into searchable intelligence, without adding a single person to the monitoring team.
Results
- Real-time PPE compliance detection across warehouse and corporate office
- Automatic alerts when equipment sits idle, surfacing lost productivity before it compounds
- Hours of video footage searchable in seconds via natural language
- Continuous anomaly detection running 24/7 without manual oversight
My Role
- Solution architect & lead engineer, system design, model/pipeline selection, GPU infrastructure, integration with on-site cameras, deployment.
Core Tech
- NVIDIA VSS Blueprint
- VILA
- NVIDIA NIM
- H100 GPU
- Computer Vision Pipelines
Footprint
- Corporate office + warehouse / production floor