Yonatan Sverdlov

I have completed the requirements for a PhD in Mathematics at the Technion – Israel Institute of Technology, advised by Prof. Nadav Dym. My research focuses on Graph Neural Networks (GNNs), geometric deep learning, learning on model and weight spaces, expressivity, oversquashing, invariance/equivariance, and Lipschitz-stable representations. I combine a strong mathematical background with strong technical skills in machine learning research, implementation, and experimentation.

I am seeking full-time Research Scientist, Research Engineer, and Applied Scientist positions.

Email / CV / GitHub / Google Scholar / LinkedIn

Publications

	Finite Probes Suffice: Identifiability and Universality for Weight-Space Learning Soutrik Sarangi, Yonatan Sverdlov, Adir Dayan, Haggai Maron, Nadav Dym arXiv preprint, under review (2026) Paper / Code
	When and How to Canonize: A Generalization Perspective Yonatan Sverdlov, Benjamin Friedman, Snir Hordan, Nadav Dym NeurIPS 2026 Paper / Code
	Monotone and Separable Set Functions: Characterizations and Neural Models Soutrik Sarangi, Yonatan Sverdlov, Nadav Dym, Abir De NeurIPS 2025 Paper / Code
	Short-Range Oversquashing in Graph Neural Networks Yaaqov Mishayev, Yonatan Sverdlov, Tal Amir, Nadav Dym LoG 2025 (Best paper) Paper / Code
	FSW-GNN: A Bi-Lipschitz WL-Equivalent Graph Neural Network Yonatan Sverdlov, Yair Davidson, Nadav Dym, Tal Amir LoG 2025 Paper / Code / BibTeX
	On the Expressive Power of Sparse Geometric MPNNs Yonatan Sverdlov, Nadav Dym ICLR 2025 Paper / Code / BibTeX
	Revisiting Multi-Permutation Equivariance through the Lens of Irreducible Representations Yonatan Sverdlov, Ido Springer, Nadav Dym ICLR 2025 Paper / Code / BibTeX

Teaching

GTIIT (Guangdong Technion–Israel Institute of Technology)

Lecturer and Tutor.