23.09.2026 aktualisiert
100 % verfügbar
Data Scientist & Machine Learning Engineer | Time-Series & Biosignals for Health | Computer Vision
Oslo, Norwegen
Nur Remote
PhD, Medical Technology (Explainable AI), NTNU, Norway, 2025Über mich
Machine learning engineer in Oslo with a PhD in explainable AI. I build models on time series, biosignals and images in Python and PyTorch, and I package them with FastAPI, Docker and ONNX. Before the PhD I spent five years as a hardware engineer. Available now for remote contracts.
Skills
Künstliche IntelligenzComputer VisionHardware-DesignMachine LearningOpenCVTensorFlowZeitreihenanalyseData SciencePyTorchDeep LearningFastAPIScikit-learnTesting (Hardware)Docker
Computer Vision
Expertise in object detection, image classification, pose estimation, synthetic data augmentation, and adversarial robustness, with hands-on project experience using YOLO, OpenCV, and PyTorch.
Time-Series and Sensor Signal Processing
Proficiency in sequence modeling, graph neural networks, human activity recognition, sensor data pipelines, signal acquisition, robustness, and perturbation testing for high-stakes domains.
Explainable AI
Deep knowledge of CAM, Grad-CAM, perturbation analysis, evaluation design, validation strategy, and explainability metrics, applied in clinical and regulatory contexts.
Tools and Frameworks
Hands-on experience with PyTorch, TensorFlow, scikit-learn, OpenCV, YOLO, ONNX, pandas, NumPy, and MATLAB for building and evaluating machine learning models.
Engineering and Deployment
Practical skills in Python, FastAPI, Docker, Git, pytest, edge and cloud deployment, and drift monitoring for reliable production systems.
Sensors and Instrumentation
Background in oscilloscope and lab instrumentation, I2C, SPI, UART, embedded C, Arduino, Raspberry Pi, and PCB bring-up and testing from a prior hardware engineering career.
Graph Neural Networks
Application of spatio-temporal graph convolutional networks for human movement signal analysis and skeleton-based human activity recognition on large 3D graph time-series datasets.
Hardware Design and QA
Experience in circuit schematic design, PCB bring-up and testing, debugging, QA-protocol development for power-supply systems, and long-term reliability testing.
Expertise in object detection, image classification, pose estimation, synthetic data augmentation, and adversarial robustness, with hands-on project experience using YOLO, OpenCV, and PyTorch.
Time-Series and Sensor Signal Processing
Proficiency in sequence modeling, graph neural networks, human activity recognition, sensor data pipelines, signal acquisition, robustness, and perturbation testing for high-stakes domains.
Explainable AI
Deep knowledge of CAM, Grad-CAM, perturbation analysis, evaluation design, validation strategy, and explainability metrics, applied in clinical and regulatory contexts.
Tools and Frameworks
Hands-on experience with PyTorch, TensorFlow, scikit-learn, OpenCV, YOLO, ONNX, pandas, NumPy, and MATLAB for building and evaluating machine learning models.
Engineering and Deployment
Practical skills in Python, FastAPI, Docker, Git, pytest, edge and cloud deployment, and drift monitoring for reliable production systems.
Sensors and Instrumentation
Background in oscilloscope and lab instrumentation, I2C, SPI, UART, embedded C, Arduino, Raspberry Pi, and PCB bring-up and testing from a prior hardware engineering career.
Graph Neural Networks
Application of spatio-temporal graph convolutional networks for human movement signal analysis and skeleton-based human activity recognition on large 3D graph time-series datasets.
Hardware Design and QA
Experience in circuit schematic design, PCB bring-up and testing, debugging, QA-protocol development for power-supply systems, and long-term reliability testing.
Erfahrung
Berufserfahrungmehr als 10 Jahre
Freelancer seit2026
Sprachen
EnglishverhandlungssicherNorwegianGrundkenntnisse
Projekthistorie
Building custom machine learning systems on time-series, sensor, and vision data for high-stakes domains, engineered for reliable production and on-device edge deployment. Pairing models with rigorous evaluation and explainable outputs that stand up to clinical and regulatory review.
Built and validated deep learning models on spatio-temporal human movement signals including pose-estimated infant motion sequences analyzed with spatio-temporal graph convolutional networks in clinical collaboration with St. Olavs Hospital. Identified limb velocity as a candidate early biomarker for cerebral palsy. Developed and validated explainable-AI methods (CAM, Grad-CAM, perturbation analysis) using targeted signal perturbation to test model reliability and robustness. Extended the evaluation approach to skeleton-based human activity recognition on a large 3D graph time-series dataset.
Led debugging, testing, and QA-protocol development for power-supply systems including a medical-device power supply. Captured and analyzed signal measurements with oscilloscopes, electronic loads, and function generators; ran thermal cycling and long-term reliability testing.