Finn Behrendt

• Led industry collaborations, delivering research-based AI prototypes for radiology • Designed CNN and Transformer models for classification, detection, segmentation, and regression • Built generative models (Diffusion) for controlled image synthesis • Used NLP to generate radiology reports

• Developed and deployed efficient deep learning models on embedded vision systems • Prototyped real-time inference pipelines using TensorRT and Embedded Linux • Delivered edge AI demos showcased at major industry exhibitions • Contributed across the full software lifecycle in an agile, cross-functional CV team