Max Ryabinin


I am a Distinguished Research Scientist at Together AI, working on efficient deep learning methods and leading the development of the training infrastructure used in multiple Together products. Previously, I was a Senior Research Scientist at Yandex Research, and in July 2023, I obtained my Ph.D. on decentralized deep Learning at the HSE University. In 2021–2022, I served as the chair of the Engineering and Scaling working group at the BigScience Workshop. Earlier, I was a Machine Learning intern at Replika and Yandex Translate.

Aside from my research work, I also teach Efficient Deep Learning Systems at HSE University and Yandex School of Data Analysis (all materials are in English) and Deep Learning at HSE University (mostly in Russian).

I am mainly interested in solving new problems that involve natural language and making latest advances in deep learning more efficient and broadly accessible. In particular, my research on decentralized deep learning has served as a basis of Hivemind, a PyTorch library for training neural networks over heterogeneous, unreliable and poorly connected hardware. I also worked on topics like uncertainty estimation in machine translation, graph-based word representations, neural networks with adaptive computation, and gradient optimization of decoding hyperparameters in language generation.

If you have any questions about my work, have ideas for collaboration or simply want to say hi, feel free to drop me a line using any of the links below!