George Dasoulas

George Dasoulas is a Postdoctoral Researcher at Harvard University's Department of Biomedical Informatics, affiliated with the Zitnik Lab. He holds a PhD in Computer Science from École polytechnique in Paris, France, and previously worked at Huawei Technologies France. His research focuses on graph machine learning, particularly in biomedical applications and telecommunications, with contributions to graph neural networks (GNNs), attention mechanisms, and topological deep learning.

Education: Ph.D., Computer Science (DaSciM group, LIX, École polytechnique); Diploma in Electrical & Computer Engineering (National Technical University of Athens). His work includes developing Lipschitz-normalized attention layers, parametrized graph shift operators, and modularity-aware graph autoencoders. He has been recognized with the 2022 Wojcicki and Troper Fellowship from Harvard's Data Science Initiative.

• Key Research Themes: Graph Representation Learning, Topological Neural Networks, Equivariant Learning, Multimodal Learning
• Applications: Biomedical Informatics, Telecommunications, Sustainable AI

His articles emphasize scalable GNN architectures, graph-based unlearning strategies, and multimodal protein phenotyping. He has contributed to open-source projects like LipschitzNorm and PGSO, and actively publishes in top conferences (ICML, ICLR, NeurIPS).