PhD Position in Physics-Informed Neural Operators for Physical Systems

This PhD position at the University of Latvia focuses on advancing neural operator methods—particularly physics-informed DeepONet and Fourier Neural Operators—to model complex physical systems governed by partial differential equations (PDEs). The project bridges applied mathematics, deep learning, and scientific computing, aiming to enhance the robustness, generalization, and physical consistency of these models for applications in hydrodynamics, crystal growth, and coastal wave dynamics. Applicants should have a strong background in applied mathematics, PDEs, and numerical methods, with prior experience in deep learning. The position is fully funded for up to three years, offering opportunities for collaboration with international research teams and additional paid involvement in related AI-for-science projects.

• Develop theoretical foundations and practical improvements for physics-informed neural operators, focusing on robustness and generalization.
• Apply neural operator methods to model complex physical systems governed by PDEs, particularly in hydrodynamics and coastal wave dynamics.
• Collaborate with international research teams to advance applications in crystal growth processes and wave simulations.
• Conduct research to enhance the physical consistency and efficiency of neural operator-based surrogate models for PDE simulations.
• Publish findings in peer-reviewed journals and present research at academic conferences.
• Engage in teaching or mentoring activities as part of the project workload.