Freelance Machine Learning Engineer building custom models for high-stakes domains: computer vision, time-series, and deep learning.
I help teams close the gap between a model that works in the notebook and one that performs reliably in production, with rigorous evaluation and explainable outputs you can defend to users, clients, and regulators.
PhD in Explainable AI, with clinical research at St. Olavs Hospital. Five years of hardware engineering before research, which is why I treat "production-ready" as an engineering claim, not a marketing one.
- PhD in Explainable AI, with clinical research at St. Olavs Hospital
- Validated at NM i AI 2026 (Norwegian AI Championship): retail shelf object detection, YOLOv8 at 1280px, 0.87 leaderboard score
- Built medical diagnosis models paired with Grad-CAM / CAM for clinician-auditable predictions
- 5 years of hardware engineering before research: PCB design, embedded systems, field application engineering
🛒 Retail Shelf Object Detection — Dense-scene detection with YOLOv8 at 1280px; validated at NM i AI.
🏥 Explainable Medical Imaging — Deep learning for clinical diagnosis with Grad-CAM/CAM explanations (PhD, St. Olavs Hospital).
📄 EU AI Act Q&A System — Production-grade RAG over the full EU AI Act with hybrid retrieval and structured evaluation.
🤖 Multi-Agent Research System — LangGraph orchestration with confidence-weighted synthesis.
Python, PyTorch, scikit-learn, OpenCV, YOLO, ONNX, pandas, NumPy, FastAPI, LangChain, LangGraph.
Specialties: Computer vision, time-series, deep learning, edge AI, Explainable AI (XAI), healthcare AI
🌐 signalsyntax.no
💼 LinkedIn: linkedin.com/in/kimjipellano
📅 Book a call with me: cal.eu/kimpellano