Browse categories
Explore
Fiverr Pro
English
$
USD
I will build a custom yolov8 computer vision and object detection solution
Ali Abdullah
About this gig
Tired of generic "AI detection" demos that fall apart on real footage? I build custom YOLOv8 object detection systems for real-world conditions tuned, evaluated, and delivered with clean code you can actually run.
I'm an AI/ML engineer specialising in computer vision. At the National Centre for Physics (Islamabad), I built Argus a multi-camera, real-time detection platform combining three YOLOv8 models (licence-plate detection with OCR, unattended-bag and weapon detection) with Flask backend, live streaming, and alert dashboard. I also built CropScan AI, a crop-disease detection app on Hugging Face Spaces.
What I build for you:
Custom YOLOv8 models trained on YOUR images and videos
Real-time pipelines OpenCV, video streams, IP cameras
OCR for licence plates, documents, labels
Preprocessing for low light, blur, small objects
Honest precision/recall reports measured performance, not promises
Flask/REST API deployment that plugs into your system
Use cases: security, retail analytics, quality inspection, plate recognition, wildlife monitoring. Send a sample of your footage I'll give you an honest assessment of what's achievable.
Get to know Ali Abdullah
Ali Abdullah
AI and Machine Learning Engineer
- FromPakistan
- Member sinceJan 2025
Languages
Urdu, English
I'm an AI & Machine Learning Engineer with experience across the full ML lifecycle, from data prep to deployment.
I build: bilingual (Urdu/English) AI chatbots and LLM assistants; computer vision systems (YOLOv8, license-plate OCR, multi-camera monitoring); ML web apps with Flask and Streamlit.
Recent work: Argus, a real-time threat detection platform from my internship at the National Centre for Physics; CropScan AI for Pakistani farmers; a chatbot for a parenting platform.
Clean, documented code, honest timelines, and clear communication from first message to delivery.
My Portfolio
FAQ
I don't have labelled training data. Can you still help?
Yes. The Basic package uses pretrained YOLOv8 models (people, vehicles, common objects) with no training data needed. For custom objects, I can advise on the fastest way to build a small dataset, or we can start with the Standard package once you have images.
How accurate will the model be?
It depends on your data quality and how distinct the objects are. That's why every Standard and Premium order includes an evaluation report with precision and recall numbers — you'll see measured performance, not promises.
Can this run in real time on a camera feed?
Yes — the Premium package is built for exactly that: video streams or IP cameras with a Flask API and dashboard. Tell me your camera setup and hardware and I'll confirm feasibility before you order.
Do I own the code and model?
Completely. You receive the full source code, trained model weights, and documentation. No lock-in, no subscriptions, no hidden dependencies beyond standard open-source libraries.
