cv

Research Resident

Yandex Research

Research Residency, Moscow

• Spearheading research on asynchronous optimization for efficient model training.

Developer

Yandex

Personalization Quality R&D Group, Moscow

• Engineered a production pipeline integrating Vowpal Wabbit-based online features into a large-scale CatBoost model, achieving a 0.1% uplift in core ranking metrics (AUC, nDCG).
• Implemented a parallelized, MapReduce-style data converter that accelerated data processing by 20x and reduced model training time from hours to minutes.

Researcher

BRAIn Laboratory, MIRIAI

Moscow

• Pioneered memory-efficient zeroth-order optimization methods for fine-tuning, achieving a 50% reduction in memory footprint.
• Developed and theoretically analyzed novel stochastic and zeroth-order algorithms for decentralized optimization, validated on variety of domains.

Data Science Track (MIPT & Yandex School of Data Analysis)

Specialized curriculum in Data Science

B.S.

Moscow Institute of Physics and Technology (MIPT)

Applied Mathematics and Computer Science

Closing the Curvature Gap: Full Transformer Hessians and Their Implications for Scaling Laws

Submitted to COLT

Egor Petrov, Nikita Kiselev, Vladislav Meshkov, Andrey Grabovoy.

Sign-SGD is the Golden Gate between Multi-Node to Single-Node Learning...

ICLR

D. Medyakov, S. Stanko, G. Molodtsov, P. Zmushko, G. Evseev, E. Petrov, et al.

When Extragradient Meets PAGE: Bridging Two Giants to Boost Variational Inequalities

Conference on Uncertainty in Artificial Intelligence (UAI)

G. Molodtsov, V. Parfenov, E. Petrov, E. Grigoriy, D. Medyakov, A. Beznosikov.

Leveraging Coordinate Momentum in SignSGD and Muon:...

ICML Workshop on Tiny Titans

E. Petrov, G. Evseev, A. Antonov, A. Veprikov, P. Plyusnin, N. Bushkov, et al.

Zero Order Algorithm for Decentralized Optimization Problems

AI Journey 2024. Published in Doklady Mathematics

A. Veprikov, E. Petrov, G. Evseev, A. Beznosikov.

Shuffling Heuristic in Variational Inequalities: Establishing New Convergence Guarantees

ICOMP 2024. Submitted to Q1 Journal JOTA

D. Medyakov, G. Molodtsov, E. Grigoriy, E. Petrov, A. Beznosikov. Awarded 1st place at Neuroinformatics 2024.

Sampling of Semi-Orthogonal Matrices for the Muon Algorithm

Submitted to Q2 Journal Doklady Mathematics

E. D. Petrov, G. V. Evseev, A. V. Antonov, A. S. Veprikov, N. A. Bushkov, S. V. Moiseev, A. N. Beznosikov.

Lead Developer, ZO-Library

Engineered a comprehensive open-source library for zeroth-order (ZO) optimization, implementing a wide range of state-of-the-art algorithms. Designed with a user-friendly, torch.optim-style API for seamless integration into existing ML pipelines.

• GitHub & Publication Forthcoming

Teaching Assistant & Mentor

Mentored undergraduate students in Algorithms, Data Structures (MIPT), and Scientific Studio (Central University), leading project-based learning and providing technical guidance.

• 2024 – Present