Zhirayr Tovmasyan

profiled.jpg

I am an M.Sc. student in Computer Science at KAUST, working with Prof. Peter Richtárik. My research interests are broadly in optimization for machine learning, particularly large-scale and distributed learning.

I am especially interested in designing optimization algorithms that better capture the geometry of modern learning models, and in understanding how computational and communication constraints should shape their design at scale.

Research Interests

Optimization for Machine Learning · Large-Scale and Distributed Learning · Federated Learning

Selected Publications

  1. ICLR
    Error Feedback for Muon and Friends
    Kaja Gruntkowska, Alexander Gaponov, Zhirayr Tovmasyan, and 1 more author
    In International Conference on Learning Representations, 2026
    arXiv HTML
  2. JNVA
    Revisiting Stochastic Proximal Point Methods: Generalized Smoothness and Similarity
    Zhirayr Tovmasyan, Grigory Malinovsky, Laurent Condat, and 1 more author
    Journal of Nonlinear and Variational Analysis, 2026
    DOI arXiv HTML
  3. arXiv
    Rennala MVR: Improved Time Complexity for Parallel Stochastic Optimization via Momentum-Based Variance Reduction
    Zhirayr Tovmasyan, Artavazd Maranjyan, and Peter Richtárik
    2026
    arXiv
  4. Under Review
    Scaled Momentum Variance Reduction in Non-Euclidean Geometry
    Egor Shulgin, Zhirayr Tovmasyan, Sarit Khirirat, and 1 more author
    2026
    Manuscript under review

Research & Industry Experience

CISPA Helmholtz Center for Information Security
Research Intern, June–August 2026
Supervised by Prof. Sebastian Stich.
Research on optimization methods for distributed learning, including FedMuon and SCAFFOLD.

King Abdullah University of Science and Technology (KAUST)
Visiting Student, September 2024–March 2025 · Remote Visiting Student, April–July 2025
Center of Excellence for Generative AI. Supervised by Prof. Peter Richtárik.

Aerodynamics LLC
Junior ML Engineer, October 2023–September 2024
Worked on computer vision projects.

Education

King Abdullah University of Science and Technology (KAUST)
M.Sc. in Computer Science, 2025–Present

Yerevan State University
B.Sc. in Computer Science, 2018–2024
GPA: 19.3/20

Selected Presentations

  • ICLR 2026 — Poster Presenter, Rio de Janeiro, Brazil
  • KAUST Rising Stars in AI Symposium 2026 — Poster Presenter
  • Distributed Training in the Era of Large Models, 2025 — KAUST