Summary
Bill Psomas is a Marie Skłodowska-Curie Postdoctoral Fellow at the Visual Recognition Group of Czech Technical University in Prague, specializing in representation learning for computer vision with nine years of research and engineering experience. His work spans image classification, self-supervised learning, deep metric learning, composed image retrieval, and open-vocabulary segmentation—now centered on RAVIOLI, a retrieval‑augmented vision-language approach for fine-grained semantic localization. He combines academic rigor (PhD in Deep Learning for Computer Vision) with practical engineering from roles in industry and research labs, having developed object detection, action recognition, and remote sensing solutions. Bill also teaches advanced computer vision and remote sensing across multiple MSc programs, bringing real-world projects and Earth-observation applications into the classroom. Notably, his background includes contributions to novel CNN and Vision Transformer architectures at IARAI and finalist work with ESA on creative Copernicus-based applications, reflecting a flair for applied, cross-disciplinary AI.
9 years of coding experience
7 years of employment as a software developer
Master of Science, Data Science and Information Technologies, Master of Science, Data Science and Information Technologies at National Kapodistrian University of Athens
Doctor of Philosophy - PhD, Deep Learning for Computer Vision, Doctor of Philosophy - PhD, Deep Learning for Computer Vision at National Technical University of Athens
English, German, Italian, French