Industrial AI

AVIS

AVIS deploys computer vision defect detection at the production line — YOLO-based real-time inference combined with Vision-Language Models for complex defect characterisation, embedded in a full ISO 9001 quality management workflow.

AVIS illustration

Key Features

Edge AI inference, standards-compliant quality management, and real-time defect detection — purpose-built for the factory floor.

🔬

YOLO + VLM Detection

YOLOv11 for real-time defect bounding box detection at <50ms per frame. Vision-Language Models (Qwen-VL, LLaVA) for defect characterisation when the anomaly cannot be classified by detection alone. Confidence thresholding with human-in-the-loop escalation.

📷

Managed Camera Fleet

ONVIF-compliant camera provisioning, health monitoring, and remote configuration. Automated lens cleaning alerts. Camera placement optimization tool. Edge AI inference on NVIDIA Jetson Orin NX for lowest latency.

📊

SPC & Quality Suite

Statistical Process Control with X-bar, R-chart, and CUSUM. FMEA failure mode database with risk priority numbers. CAPA (Corrective and Preventive Action) workflow. 8D problem-solving template with root cause analysis tools.

🏭

ISO 9001 / IATF 16949

Quality management system designed to ISO 9001:2015 requirements. IATF 16949 automotive extensions — PPAP (Production Part Approval Process), Control Plans, MSA (Measurement System Analysis). Customer-specific requirements (Ford Q1, GM BIQS) templates.

💊

21 CFR Part 11 (Pharma)

Electronic records and signatures compliant with FDA 21 CFR Part 11. Audit trail meets GxP requirements. CSV (Computer System Validation) documentation package included. Used in pharmaceutical and medical device manufacturing.

Edge AI Deployment

Models run at the edge — no cloud dependency for inference. NVIDIA Jetson Orin NX (100 TOPS) standard deployment. Docker-based model updates pushed from central management. Federated learning across sites with privacy preservation.

Technical Specifications

AI Performance

  • <50ms per frame inference (Jetson Orin)
  • YOLOv11 + Qwen-VL models
  • mAP@0.5 > 0.92 on internal benchmarks
  • Retraining pipeline with active learning

Quality Standards

  • ISO 9001:2015
  • IATF 16949:2016 (automotive)
  • 21 CFR Part 11 (FDA pharma)
  • AIAG FMEA 4th Edition

Hardware

  • NVIDIA Jetson Orin NX (standard)
  • ONVIF IP cameras (any brand)
  • Industrial ruggedized enclosures
  • PLC integration via OPC-UA

Integration

  • SAP QM module integration
  • REST API for MES/ERP
  • MQTT for IoT sensor fusion
  • Grafana dashboards out of box

Deploy AVIS on your production line

Contact our team to discuss your requirements. We respond within 24 hours.