- Participated at 5 workshop and 4 full papers accepted to the A* conferences (ICML, ICLR, NeurIPS).- First author paper on SAC-RND algorithm (accepted to the ICML 2023) was covered by all major Russian news media. - Was awarded with Yandex ML Prize (Ilya Segalovich Award) in the First publication nomination.- Developed and maintained the CORL library, which currently has over 1k stars on the GitHub and has become one of the most popular libraries for offline reinforcement learning. Technical paper describing the main results and philosophy of the library has been accepted at NeurIPS 2023.- Actively attended meetups and conferences, presenting research accomplishments. Presented CORL library at PyCon Russia 2023 (rated top 8 out of 25+ by the attendees). Presented CORL & SAC-RND at the meetups organized by Tinkoff. Gave over 8 talks on my research to university students from three cities (Moscow, St Petersburg, Kazan). Gave several interviews about my work to Forbes Russia, Tinkoff Magazine, vc.ru and others.

- Conducted research on offline reinforcement learning. Publicly presented the results of the internship at the Innopolis University.- Proposed and verified more than 5 research proposals during the internship. One of them was accepted to the 3rd Offline Reinforcement Learning Workshop at NeurIPS 2022.

- Worked on a Improving Upside Down RL coursework. Developed most of the infrastructure for conducting experiments and training in PyTorch. Tested various modifications of the Soft Actor-Critic algorithm in the Upside Down RL setting.- Participated in the MineRL NeurIPS 2021 competition, taking 5th place as well as receiving a $200 award for the community support nomination. As a result, our approach was described in the paper accepted to NeurIPS 2021.

- Developed most of the infrastructure for conducting experiments and training in PyTorch. Designed and performed the main experiments. Achieved an average threefold improvement over baselines on a variety of metrics. Results were published in the Photonics and Nanostructures journal.

- Wrote a pipeline for the training of graph embeddings using ALS (on millions of nodes) and top-n recommendations of the most similar users using distributed kNN search. Results were tested with an A/B test.- For the correct sampling of the graph and the tuning of the hyperparameters, also wrote my own implementation of the Metropolis-Hastings random walk on pyspark.