Muhammad R.

I build computer vision and video analytics that turn your existing cameras into real-time alerts. 🎯 Object Detection | Object Tracking | Video Analytics | AI Visual Inspection | Edge AI CCTV & Surveillance AI | Retail AI | Manufacturing QC | Healthcare Imaging AI You already have cameras, and a clear problem you want them to solve. Maybe it's intruders, missing PPE, long queues, or defective parts. The hard part is getting AI that's accurate on your footage, fast enough for live video, and stable when lighting, angles or crowds change. Generic models give false alarms. Slow inference misses events. And many pilots never get past one camera. That's where most projects get stuck. I design and deploy real-time computer vision systems end to end. I start with your own footage and go through training, optimization and deployment to live alerts and dashboards your team actually uses. NVIDIA Jetson | TensorRT | ONNX | DeepStream | AWS | Azure | Docker | FastAPI | RTSP Streams Training a model is only half the job. What you get from me: ➤ Custom models trained on your camera footage, not just public datasets ➤ Edge deployment on NVIDIA Jetson with TensorRT and FP16/INT8 optimization ➤ Multi-camera tracking with zone, time and duration rules to cut false alarms ➤ Integration with your VMS, POS, MES/PLC or PACS systems ➤ Instant alerts, dashboards and reports your team can act on Key Accomplishments (projects I've led at SwiftVision AI): ⭐ Security: real-time CCTV analytics across 47 facilities, 98% fewer incidents, 50% faster response ⭐ Retail: footfall and queue analytics across 220 stores, +95% conversion, 42% shorter queues ⭐ Manufacturing: inline defect detection across 14 plants, 90% fewer defect escapes at sub-80ms ⭐ Healthcare: AI radiology triage across 8 imaging sites, 90% faster triage on 12k+ studies a month ⭐ 7+ years building applied computer vision, from model training to edge deployment Domain Expertise: ✅ intrusion detection | restricted zones | loitering | crowd density | face blurring | vehicle theft ✅ PPE compliance | forklift safety | hygiene monitoring | staff monitoring | workplace safety ✅ defect detection | PCB inspection | conveyor QC | barcode and print verification | measurement ✅ people counting | footfall | heatmaps | queue and dwell time | shelf monitoring ✅ grain grading | fruit counting | packaging alignment | production line monitoring ✅ tumor segmentation | MRI analysis | medical image segmentation | radiology triage Tech Stack: YOLO (v8, 11), RT-DETR, Detectron2, SAM, ByteTrack, DeepSORT, PyTorch, OpenCV, TensorRT, DeepStream, ONNX, NVIDIA Jetson, RTSP, FastAPI, Docker, AWS, Azure, DICOM / PACS 💬 If you have cameras, video or images and need AI that is accurate, fast and actually deployed on site, send me your use case and a short sample clip. I'll tell you what's achievable and how I'd build it.