Remote: Yes, part-time (contract or advisory)
Willing to relocate: No
Technologies: Python, C++, SQL, PyTorch, TensorFlow/Keras, Transformers, Hugging Face, YOLO, Segment Anything, LightGlue, OpenCV, scikit-learn, ONNX, FAISS, Kalman filters, OCR (Tesseract, docTR), MLflow, W&B, Ray, Docker, AWS, GCP, Azure
Résumé/CV: https://bit.ly/cvhackernews
Email: kujtim.rahmani [at] gmail [dot] com
Computer vision and ML engineer with a PhD in computer science (Germany) and 12+ years building vision systems from data collection to production.
What I've built: real-time smoke and fire detection (day and night), multi-camera crowd tracking, object detection, counting and 3D reconstruction in warehouses, defect detection in X-ray and CT scans of manufactured parts (99.9% accuracy on pore detection), visual search with metric learning and FAISS, image-based oral health diagnostics, and OCR/document extraction pipelines.
Teams: led a computer vision team of engineers and annotators through the full cycle, from data acquisition and annotation to deployment. Also led a 26-month detection rebuild that reached real-time performance and saved millions of $
Advisory and contract: AI architecture for a national smart-city program (urban detection, counting, crowd tracking), a museum digitization platform (segmentation and visual search), and distributed training infrastructure for an autonomous driving project. I've also advised AI projects in fintech and manufacturing.
Teaching: university lecturer in Linear Algebra and Data Science. I've trained 2,000+ public-sector and banking staff in applied generative AI (Microsoft Certified Trainer).
Looking for: part-time contract or advisory work, including hands-on CV/ML development, architecture reviews, or helping a team take a model from prototype to production.