As no system is the same, for any engineer, the learning process should be at least as important as any existing knowledge.
I built a mobile app that detects and categorizes Vienna convention traffic signs. The detector uses YoloX, while the classifier uses a lightweight EfficientNet model. The training data was sampled across multiple public traffic sign datasets, and is refined via a complex image augmentation pipeline. Training and Inference is done in PyTorch. The inference server communicates with the mobile application via WebRTC.
View on GitHubI worked on software that allows Drone Field Testing in collaboration with the QSE institute at TU Wien.
View IEEE paperI built the websites for the synthetic material producer TTPH, and for Bio-Hof Binder.
Explore ttph.eu Explore bio-hof.atI made a TicTacToe AI in PyGame based on the minimax algorithm with an optional difficulty where you cannot lose.
Play