Machine Learning engineer on an Object Detection system for road damage detection. Worked on dataset improvement and set up a scalable workflow for dataset iteration. Maintained Kubeflow pipelines for training and batch inference, ML-related tooling for dataset management, version tracking and customer onboarding. Example: I developed a dataset versioning system on top of Dolt. I led a team of three people (including me) and handled communication with stakeholders.

Applied ML theory and causality to Object Detection. Implemented out-of-distribution learning methods from literature and proposed by myself in BrainCreators’ Detectron2 fork.

Part-time position in a team of three students. We designed and produced a usable prototype battery (690V, 12kWh) for a deep-sea (4 km) application, without a pressure housing. My main responsibilities were the custom design of the electronics and software, and general project management.

Full-time position, leading the electrical design and software development for a Formula Student race car. This longtime high-ranking team has been world champion five times. Our car won FS Netherlands and was highly competitive in two other competitions. I managed a team of 10-15 people, made high-level conceptual choices, and was directly involved in software and hardware development. See video here.