Research
My group develops scalable algorithms for PDE simulation by combining randomized numerical linear algebra with hierarchical structure and performance-aware implementation. I’m especially interested in methods that are both theoretically grounded and practical on modern architectures.
Topics
- Randomized compression and sketching
- Hierarchical matrix methods and fast solvers
- Preconditioning and scalable implementations
Current directions
- Fast solvers for large-scale PDEs
- Robust approximations with error control
- GPU-friendly kernels and parallel scalability
News
- Mar 2026 Serving on the program committee for ISC 2026 in the Algorithms & Performance area/track.
- Dec 2025 New preprint: “A fast spectral overlapping domain decomposition method”.
- Aug 2025 Joined the faculty at The Ohio State University. Go Bucks!
- Jun 2025 Plenary talk at the Householder Symposium XXII on hierarchical and randomized solvers.
Prospective Students
I’m recruiting students at multiple levels.
- PhD and MS students: Please feel free to reach out if you are interested in working with me.
- Undergraduate students at OSU: Strong undergraduates interested in research in numerics are welcome to reach out. Helpful preparation includes coursework in linear algebra, real analysis, ODEs, and PDEs, along with some programming experience.
- Getting in touch: If you think our interests might align, please feel free to email me with your CV and a brief note about your background and research interests.
Selected Publications
Contact
Email: yesypenko.1@osu.edu
Office: MW 740
You can also find my CV and publication list via the links above.