Georgios Kissas
Senior Data Scientist, Research, Swiss Data Science Center
I am a Senior Research Data Scientist at the Swiss Data Science Center of ETH Zurich, working on AI for science. My current research focuses on generative models for extreme-scale Lagrangian simulations of subsonic turbulence, verifiable scientific discovery, and verification-driven AI agents. I also develop geometry-aware neural operators and physically faithful representations of scientific data, including methods for modeling transport, learning on irregular geometries, and discrete tokenization.
Previously, I was a Postdoctoral Researcher at the ETH AI Center, working with Siddhartha Mishra, Eleni Chatzi, and Laura De Lorenzis.
I received my PhD from the University of Pennsylvania, advised by Paris Perdikaris. During my PhD, I worked on neural operators, generative modeling, and inverse problems, with applications to cardiovascular biomechanics and fluid dynamics.
Before that, I studied Computational Mechanics at the National Technical University of Athens, where I worked with Doros Theodorou, and Mechanical Engineering at the University of Patras.
news
| Oct 20, 2026 | Our paper Rózsa: Certified Equational Reasoning is now available on arXiv! We placed 6th overall on the GPT-OSS-120B leaderboard and 8th on the Gemma-4-31B-IT leaderboard in Stage 2 of the SAIR Mathematics Distillation Challenge (Equational Theories). |
|---|---|
| Oct 08, 2026 | Our papers Advectra, METRO, and Phaedra have been accepted at NeurIPS 2026. |
| Sep 17, 2026 | I was invited to serve as an Area Chair for ICLR 2027. |
| Jul 17, 2026 | Our paper Neuro-Symbolic AI for Analytical Solutions of Differential Equations was presented as an oral contribution at the Greeks in AI 2026 Symposium. |
| May 24, 2026 | I am going to present as an invited speaker to PhysML Workshop 2026 in Oslo. Check it out! |
selected publications
- Neuro-Symbolic ODE Discovery with Latent Grammar Flow2026
- Phaedra: Learning High-Fidelity Discrete Tokenization for the Physical Science2026
- Neuro-Symbolic AI for Analytical Solutions of Differential EquationsIn Forty-Third International Conference on Machine Learning, 2026
- The Language of Hyperelastic MaterialsComputer Methods in Applied Mechanics and Engineering, 2024
- Learning Operators with Coupled AttentionJournal of Machine Learning Research and Thirty-sixth Conference on Neural Information Processing Systems, Journal to Conference Track, 2022