Alexander Gaponov (Aliaksandr Haponau)

Hey! I’m Alexander (you can call me Alex or Sasha). I am currently pursuing my Master’s degree at King Abdullah University of Science and Technology (KAUST) in the research group of Peter Richtárik.

My current research interests are primarily focused on distributed optimization and AI safety.

Publications

• Error Feedback For Muon And Friends. Kaja Gruntkowska, Alexander Gaponov, Zhirayr Tovmasyan, Peter Richtárik. ICLR 2026.

• Pay Attention to Attention Distribution: A New Local Lipschitz Bound for Transformers. Nikolay Yudin, Sergei Kudriashov, Alexander Gaponov, Maxim Rakhuba. ECCV 2026.

• Byzantine-Robust and Private Federated Optimization under Weaker Assumptions. Rustem Islamov et al. UAI 2026.

• Communication-Efficient Gluon in Federated Learning. Xun Qian, Alexander Gaponov, Grigory Malinovsky, Peter Richtárik. 2026.

Work Experience

• Visiting Student Research Program (KAUST, Optimization in Machine Learning, May 2025 – August 2025)
  • The main result: “Error Feedback For Muon And Friends” paper.
• Intern ML Engineer (Yandex, Neuro Basic Quality Model Development Group, June 2024 – October 2024)
  • Developed Multi-rephrase pipeline components for the Neuro LLM-search service, achieving significant growth in usefulness metrics for complex and quiz-style queries.
  • Trained, distilled, and fine-tuned text classification and generation models using Reinforcement Learning techniques.
• Intern ML Engineer (Yandex, Alice (smart assistant), Product development and quality improvement group, March 2023 – August 2023)
  • Transitioning from Boosting to LLMs for scenario classification, significantly enhancing classification quality for long-context queries.

Education

• M.S. in Computer Science (KAUST, 2025-2027)
  • GPA 3.89/4
• B.S. in Applied Mathematics and Information Science (HSE, 2021-2025)
  • Graduated with Honors, GPA 8.81/10

Presentations

• UAI 2026 (Poster: Byzantine-Robust and Differentially Private Federated Optimization under Weaker Assumptions, August 2026)
• ICLR 2026 (Poster: Error Feedback For Muon And Friends, April 2026)
• Workshop on Distributed Training in the Era of Large Models, KAUST (Poster: Error Feedback For Muon And Friends, November 2025)
• Spring School on Machine Learning, HSE (Poster: Matrix Exponential for Parametrizing Orthogonal Convolutions, May 2024)

Teaching Experience

• Teacher Assistant (Deep Learning in Audio) (HSE, Faculty of Computer Science, 2025)
• Teacher Assistant (Matrix Computations) (HSE, Faculty of Computer Science, 2024)