Governing-equation-driven models for fluid dynamics, ballistics, structural response, and coupled systems.
PINNs, reduced-order models, and hybrid solvers that embed physics directly into learning architectures.
Scalable numerical implementations optimized for modern compute architectures.