Browse categories
Explore
Fiverr Pro
English
$
USD
I will troubleshoot and fix your computer vision python scripts
Djalil K
About this gig
Stuck with a "CUDA Out of Memory" or a cryptic YOLO error? Im here to save your project.
Don't waste hours fighting with StackOverflow or generic AI suggestions. As a Computer Vision engineer, I specialize in resolving complex issues that AI assistants can't solve.
What I can fix for you:
- Model Training Errors: Fixing mAP at 0%, loss issues, or YAML config bugs.
- Hardware & Environment: CUDA/cuDNN setup, PyTorch/TensorFlow version conflicts.
- Export & Deployment: Debugging ONNX, TensorRT, or OpenVINO conversion errors.
- Logic & Performance: Fixing Object Tracking IDs, FPS drops, or custom CV pipelines.
FREE INITIAL CHECK: Send me your error log or a snippet of your code. I will tell you if I can fix it before you even place an order!
Fast delivery, clean code, and explanations included so you don't get stuck again.
Message me now and let's get your code running!
Get to know Djalil K
Djalil K
Computer Vision Engineer
- FromFrance
- Member sinceApr 2026
Languages
English, French
Hi, I’m Djalil!
I am a Machine Learning Engineer with over 5 years of experience specializing in Computer Vision and Deep Learning. I help businesses and startups build custom AI solutions that turn visual data into actionable insights.
What I do:
Custom Object Detection (YOLO expert)
Image Segmentation & Tracking
End-to-end AI Pipeline development
Python & PyTorch integration
My goal is to deliver high-quality, production-ready code tailored to your specific needs.
Ready to unlock the power of AI? Send me a message, and let’s discuss your project!
My Portfolio
FAQ
Which YOLO versions do you support?
I support all major versions. Whether it's training, inference, or model export, I can fix issues across the entire YOLO ecosystem.
Can you fix CUDA or GPU-related errors?
Yes. I can help resolve "CUDA Out of Memory" issues, driver incompatibilities, and PyTorch/TensorFlow environment conflicts to ensure your model runs correctly on your hardware.
Do I need to send my entire dataset?
Usually, no. A small sample of your data (5-10 images) and your configuration file are typically enough for me to reproduce the bug and find a solution.
What if you can't fix the bug?
My priority is your satisfaction. If I determine that the issue is outside my area of expertise or technically impossible to solve during our initial chat, I will let you know immediately. If I can't deliver a fix, you get a full refund.
Will you explain how you fixed the error?
Absolutely. I provide comments in the code and a brief explanation so you can understand the "why" behind the fix and avoid the same mistake in the future.
