In this session, we'll discuss our approach to applying deep learning algorithms to full car inspection and address the challenges involved in real time data processing. We plan to demonstrate how we apply convolutional neural networks used for geometry verification during the complete vehicle inspection on BMW assembly line. The complete inspection of a fully assembled car involves verifying the geometry of the car, ensuring the correct customization, and detecting possible defects. AI-aided computer vision eliminates problems of traditional approaches arising from highly reflective surfaces and specular reflections. Using deep learning models and the processing power of a multi-GPU system, we can make the inspection happen in real time.