Guillaume Stern

-instructed computer vision classes to adults in continuing education

-dealt with theoretical aspects of neural networks, deep learning, CNN's, transformers

-designed and implemented Python pipelines, fine-tuned CNN and transformer models for: -> trademark classification: https://vienna-assistant.branddb.wipo.int/

-> image segmentation: extracting figures and captions from patent drawings

-designed and implemented Python pipelines using deep learning models for: -> object detection: tracking top-level skiers in competition videos to improve their performances -> image segmentation: estimating the solar potential of buildings in Geneva

https://ww2.sig-ge.ch/vitale_vert/projets/cadastre-des-facades-solaires