Laurent Kalpers

Design and deployment of deep learning models for real-time object detection and tracking on embedded devices (e.g. Jetson, NXP). I work on the full lifecycle: model prototyping, training, optimization (ONNX, TensorRT, TensorFlow Lite), and integration in performance-constrained environments. Projects include human detection, segmentation, and synthetic data generation for defense-grade systems.

Worked on data science and engineering projects for major clients in the banking and energy sectors. Maintained and improved data pipelines in complex Azure-based environments, developed predictive models for energy consumption, and contributed to a fiber rollout optimization tool using clustering and geospatial data. Strong collaboration with business and data teams.

Developed a mobile and wearable system to detect stress from physiological signals (smartwatch + Android app). Collected and labeled real-world data, extracted features, and trained ML models (SVM, XGBoost). Collaborated with researchers at UMons for algorithm refinement.