Dmitry Kabanov

Software Engineer in the MaRDI Open Interfaces project (MaRDI = Mathematical Research Data Initiative, Germany-wide project for IT services for mathematicians). I am responsible for defining software architecture and code development in this project. Stack: C (core libraries); Python & Julia for numerical libraries; CMake for software building; Google Test & Pytest for unit testing.

Trained and test classification neural networks for NLP projects (extraction of information from business documents). - Worked in a team of 5 people, with responsibility on three neural networks. - Set up job-queueing system to facilitate organized team work on shared computational resources - Stack: Python, TensorFlow 2, Keras, Weights&Biases

Applied physics-informed neural networks for problems in fluid dynamics. Physics-informed neural networks = supervised learning for neural networks with added information about physical processes behind the data.