Publications
The common thread across these papers is not a single method family but a repeated objective: make complex PDE-based workflows more computationally workable, more scalable, and easier to move toward actual use.
Featured
This paper addresses one of the harder deployment barriers in reduced-order modelling: geometric variation. The approach combines unfitted methods with reduced basis ideas so parameterized geometries can be handled without the usual remeshing burden.
Theme 1
Papers in this track deal with the cost of approximation spaces, compression, hyper-reduction, and implementation choices that matter once problem sizes become real.
Theme 2
The software paper is included here not as documentation of a codebase, but as part of the technical argument for making methods accessible and extensible.
2025
Turns reduced-order techniques into a reusable software layer with an emphasis on interpretability, extensibility, and computational efficiency in Julia.
2025
Shows how tensor-train structure can reduce basis construction and operator approximation costs without giving away too much robustness.
2024
Develops space-time interpolation ideas for transient reduced models where the cost of residual and Jacobian handling is usually the limiting factor.
2024
Extends space-time reduced basis approaches to unsteady flow settings and provides part of the groundwork for later transient-model work.