Denis Uzhva Email & Phone Number

• Developing and supporting recommender services for audio streaming.
• Full pipeline from experiments to scheduled automatic data preparation, model training (Airflow) and service deployment (FastAPI).
• Primarily focused on classical recommender systems (ALS, boosting) for automatic music playlist generation, also open to implement more complex approaches, such as transformers.Key achievements: developed and deployed.

St Petersburg, St Petersburg City, Russia

• Developed deep learning and noise simulation models for night photography enhancement.
• Investigated and implemented machine learning compression and reconstruction algorithms for geophysical data encoding.
• Developed reinforcement learning models for solid body topology prediction, based on desired body durability.
• Conducted statistical A/A and A/B tests for the developed machine learning systems.Key achievements: the investigated and developed ML/DL systems outperform conventional and state-of-the-art approach by a significant.

St Petersburg, St Petersburg City, Russia

• Implemented state-of-the-art machine learning methods for data analysis in the CERN physical experiments.
• Responsible for design of architecture and software implementation of machine learning models, as well as data augmentation, feature generation and visualization.Key achievement: the developed models demonstrate.

St Petersburg, St Petersburg City, Russia

• Developed mathematical theory for efficient multi-agent networks control.
• Responsible for the development of the software environment for network simulations.
• Developed deep learning models for efficient multi-agent system encoding.Key achievements: the developed scalable machine learning systems successfully encode big data emerging in large-scale networks by compressing it.

Doctor Of Philosophy - Phd, Mathematics And Computer Science

Master'S Degree, Computer Software Engineering

Bachelor'S Degree, Physics