I will train custom yolo object detection and computer vision models

K
kazmisohail
K
kazmisohail
Sohail A

About this gig

Need an accurate real-time Computer Vision model trained on your own camera, drone, medical or satellite imagerynot a generic off-the-shelf demo?

I am an AI & Computer Vision engineer with real-world experience developing multi-class deep learning pipelines for satellite imagery enhancement (cloud detection, haze mitigation, and shadow removal) as well as real-time parking space detection systems using YOLO and U-Net architectures.

My Computer Vision Services:

  • Custom Object Detection: Training YOLOv8/v11 models on your custom classes
  • Image Segmentation: Pixel-level segmentation using U-Net, TensorFlow, Keras, and PyTorch
  • Dataset Annotation & Augmentation: Bounding boxes and polygons prepared in RoboFlow
  • Image Processing & Geospatial AI: Noise/haze removal, OpenCV video analytics, and multi-spectral image classification
  • Deployment: Exporting weights (.pt, .onnx) wrapped in a fast Python/Flask REST API

What You Will Receive:

  • Trained model weights + confusion matrix & mAP/IoU evaluation charts
  • Clean Python inference code for images, video streams, or webcam feeds
  • Full documentation for retraining or deploying on your server

Get to know Sohail A

Sohail A

AI and Computer Vision Engineer, Python, YOLO, NLP and Flutter Developer

  • FromItaly
  • Member sinceOct 2026
  • Avg. response time1 hour
  • Languages

    English, Urdu, Italian
Ciao! I am Sohail, a Computer Science graduate, currently pursuing a Master's degree in Engineering Technology for Strategy and Security at the University of Genoa, Italy. I specialize in building production-ready Python AI systems, custom YOLO/U-Net computer vision models, automated data pipelines, and cross-platform Flutter applications. Having trained over 60 developers in data science and software engineering and built real-world systems ranging from satellite imagery deep learning models to NLP document rankers, I deliver clean, documented code with reliable CET time-zone communication.

My Portfolio