Machine Learning Research Engineer at Stadler Rail

Led end-to-end development of computer vision software for collision warning (object detection, tracking, distance estimation), enabling replacement of external supplier software costing over CHF 1M. Responsible for algorithm design, dataset compilation, model training, and deploying on edge devices via model compilation and quantization. Delivered NLP information retrieval and on-premise LLM-based document processing. Supervising ETH Zürich master's thesis on vision-based rail defect detection (publication in preparation). Led the technical evaluation and hiring process for a new ML Engineer.