Browse categories
Explore
Fiverr Pro
English
$
USD
I will develop computer vision models with pytorch, yolo object detection and ocr
Pakistan
I speak Spanish, French, German, Italian, Arabic, Russian, Chinese, Urdu
116 orders completed
PhD AI Engineer, AI Chatbots, AI Agents, ML, Flutter AI Apps, DevOps
I'm a PhD AI Engineer who builds AI chatbots, AI agents and machine learning models for businesses. As a Software Engineering lecturer with 8+ years and 100+ delivered projects, I build RAG chatbots t...
About this Gig
Need an AI model that sees and understands images or video: detecting objects, classifying defects or diseases, or reading text? I develop custom computer vision models in Python.
I'm an AI engineer and PhD scholar with 8+ years of experience in machine learning, deep learning and OpenCV.
What I build:
- Object detection and tracking (YOLOv8/YOLO11, Faster R-CNN)
- Image classification (CNN, ResNet, EfficientNet, ViT)
- Image segmentation (U-Net, SAM)
- OCR and document scanning (Tesseract, PaddleOCR, cloud OCR)
- Defect, crop disease and medical image detection
- Face and pose detection, people counting, CCTV analytics
- OpenCV image processing pipelines
You get:
- A trained model with clear metrics (mAP, accuracy, confusion matrix)
- A Jupyter notebook and inference script
- Optional API (FastAPI), mobile (TFLite/CoreML) or edge deployment
- Help with dataset preparation and labelling
Tools: PyTorch, TensorFlow, OpenCV, Ultralytics, Roboflow, Hugging Face.
Send me sample images and your goal for a free feasibility check.
My Portfolio
Other Data Science & ML Services I Offer
FAQ
How many images do I need?
About 100–300 per class for a custom model; fewer if a pretrained model fits. I'll check your data first.
Can you label my images?
Yes. Labelling help is included in Standard and Premium, with more available as an extra.
Which accuracy will I get?
I report mAP/accuracy on a held-out test set, and tell you before training if your data is too limited.
Can it run in real time on video or CCTV?
Yes, with YOLO on GPU or optimised edge models (Jetson, Raspberry Pi, mobile).
Can you deploy it in my app?
Yes: a FastAPI endpoint, a TFLite/CoreML model for mobile, or a Docker container.
Can you do OCR on invoices or IDs?
Yes, with OCR plus layout extraction into structured JSON/Excel.
