I will build custom computer vision pipelines using yolo and mediapipe
Full Stack Developer
Over deze dienst
Bring intelligent visual automation directly into your software pipeline.
Many developers can run a basic script off a local webcam, but building low-latency and production-ready computer vision applications requires clean multi-threaded backend architecture. I engineer high-performance visual AI models optimized for real-time tracking, custom model detection, and seamless edge-to-cloud deployments.
What I Can Build For You:
- Object Detection & Tracking: Custom YOLO pipelines fine-tuned to recognize and monitor specific target classes, boundaries, or entry zones.
- Human Pose & Gesture Estimation: Lightweight multi-person landmark mapping with MediaPipe for sign language, sports metrics, or posture alerts.
- Custom Dataset Training: Specialized dataset cleaning, bounding-box image filtering, and validation tracking loops.
- Stream Optimization: Multi-threaded OpenCV frame processing optimized to prevent frame lag over enterprise IP cameras.
My Vision Stack:
- Core Tech: Python, OpenCV, MediaPipe, YOLOv8/YOLOv11, PyTorch, TensorFlow, Docker
NOTE: Please message me regarding your explicit project scope and dataset availability BEFORE ordering.
API's:
Overige
Programmeertaal:
Python
Tools:
Jupyter-notitieboek
•
opencv
•
tensorflow
•
CVAT
•
Colab
•
PyTorch
Frameworks:
Scikit-learn
•
keras
•
PyTorch
Mijn portfolio
Andere Data science en ML diensten die ik aanbied
Veelgestelde vragen
Your account is new. Why should I hire you?
I am new to Fiverr, but not to the engineering industry. I have over 2 years of professional background as a backend software engineer building production-grade architecture. I launched this profile to reach global clients, giving you senior-level engineering at entry rates.
Do I need to provide a fully labeled dataset?
If you have raw video clips or images, I can help architect an automated preprocessing pipeline to structure them. For full custom model training (Standard/Premium), having raw bounding box annotations or clear folders sorted by class is ideal. Contact me to verify your assets.
Can this run on an ordinary computer or a web server?
Yes. By focusing heavily on lightweight frameworks like MediaPipe and optimized YOLO configurations, your custom model pipelines can execute smoothly on standard consumer CPUs or low-cost cloud server instances without requiring an enterprise-tier GPU array.
Do I get full ownership of the code?
Absolutely. You receive 100% ownership of the clean, fully documented script or Jupyter Notebook (.py or .ipynb). I never obscure, encrypt, or hide my engineering behind third-party software paywalls. Everything created for your pipeline is entirely yours.

