Dmitry Evdokimov Email & Phone Number

After my bachelor degree in experimental physics and working in the IceCube collaboration (we optimized the reconstruction methods of the muon flux energies in the IceCube neutrino observatory), I realized that data analysis methods (simple statistics) we used were rather poor, and wanted to go deeper — to Machine Learning and stuff. I went to the Data Science industry then, working as a junior for a small company that analyses data from online cashboxes. There I learned basic ML and tried the first DL via keras. Then I went to the Undatify startup, where we developed an e-mail classifier via BERT-like models, FastText embeddings and Muse embedding mapper. Also we experimented with deepspeed and relation extraction. Unfortunatelly, the startup was closed due to the pandemic, and I entered Kaspersky lab. My motivation was not only to do machine learning, but to create services for some social good (information security in this case). Again, I worked with transformers and some tabular ML tasks. At the same time I entered HSE master’s program on Data Science and started to do my scientific work in Lambda lab towards Hamiltonian learning: we parameterized a hamiltonian of different quantum systems via the neural ODE in order to train to reconstruct the quantum evolution for the given initial state and outer conditions. My second scientific work there is a fast simulation approach for NICA MPD: we reconstruct the detector output signal for the given event properties via the GAN. My part is trying new experimental architectures of this GAN which make it easier to train. Currently I’m looking for a research/industrial position in Machine Learning.