Kirill Tamogashev
Summary
Ph.D. student at the University of Edinburgh, focusing on diffusion-based generative modelling. My research interests include generative modelling, diffusion models, diffusion samplers, Schrödinger bridges, and probabilistic inference. Experienced in AI for audio processing, large-scale multi-GPU training, and efficient inference. I hold an MSc in Applied Mathematics and a BSc in Economics.
Industry Experience
University of Edinburgh
September 2025 - present
Samsung Research
June 2023 - January 2025
- Research engineer working on cutting-edge speech enhancement algorithms for wearable devices and laptops.
- Successfully developed SOTA fast and efficient universal speech enhancement algorithms based on adversarial and representation learning methods, published at NeurIPS 2024, [PAPER].
T-Bank
November 2021 - June 2023
- Machine learning engineer working on text-to-speech models.
- Successfully improved the performance of existing text-to-speech models by introducing a prosodic model that made the resulting voice sound 5-10% more natural (according to internal tests).
- Designed a framework for deploying and serving text-to-speech pipelines in production.
Education
Ph.D. in Informatics, University of Edinburgh
1/2025 - present
- Supervised by N. Malkin.
MSc in Applied Mathematics and Computer Science,
National Research University Higher School of Economics
9/2022 - 5/2024
BSc in Economics, National Research University Higher School of Economics
9/2018 - 5/2022
Skills
- Programming: Python (used in both research and production), PyTorch, JAX, TensorFlow, C++.
- Languages: Russian (native), English (proficient), German (B2), French (B1).