My research focuses on developing neural networks for computer vision and spatio-temporal forecasting, with primary application to environmental simulation (especially ocean hydrodynamics). I also explore methods in manifold learning for high-dimensional data analysis. I am driven by building tools that bridge advanced AI with impactful domain science.
I am the part of the NSS Lab team. My key projects include:
- TorchCNNBuilder - the Python library for modular and configurable construction of convolutional neural networks in PyTorch, designed to streamline prototyping and experimentation.
- MANUL - the research tool for data-driven manifold learning and regularization with a wide scope.
- Сontributions to the automated machine learning framework FEDOT.
- Development and implementation of solutions for the forecasting of ice conditions in the Arctic region: DL, classical models, cartography and GIS.
- Research and development for automatization of CV-pipelines for microscopy imagery.
My research papers are available on Google Scholar.