- camera_calibration - Improve camera calibration for cameras with wide FOV by simulating the distorted lattice and predicting possible locations of the previously undetected points. My Master's diploma project.
- SynRD - A Differentially Private (DP) Synthetic Data benchmarking package, made in collaboration with New York Univerity (paper).
- adversarial_attack - Generate adversarial noise using genetic algorithms to hide the text from an OCR toolchain (Tesseract) such that a human can still read it.
- smart_rockets - pathfinding in a labyrinth using genetic algorithms.
- FirstAidBot - a telegram bot and an MVP app for helping civilians delivering first aid. Rust, Flutter, CI/CD, auto-deploy to cloud.
- replicated_log - a web server, which broadcasts messages to clients. Supports quorum, health checks, message resending etc.
- rust_algs - implementation of disjoint set, segment_tree, trie, list_node and tree_node (from leetcode).
- goto-elimination - An implementation of Taming Control Flow A Structured Approach to Eliminating Goto Statements.
- RGOL - reverse game of life using genetic algorithms.
- cpp-algs - collection of useful C++ stuff, used mainly for competitive programming.
- brainfuck - An implementation of the brainfuch interpreter and debugger.
- libft - my implementation of the C standard library, including printf.
- minitalk - send strings using user signals
- sem_1
- distributed_systems: Master/client Rust/Python distributed system in docker compose with various options.
- intro_to_data_science: Group project, Ukrainian book generation. I was responsible for fine-tuning GPT-2 on the grid of authors and genres. We also used VQGAN+CLIP for image generation.
- linear_algebra: Various projects (both python and math in latex) on topics including orthogonality, eigenvalues, eigenvectors, matrix factorization and iterations and numerical optimization.
- statistics: Just a lot of math.
- warehousing: Data varehousing. ETL in sql.
- sem_2
- bayesian_methods: Statistics, Kalman filter, particle filter.
- DL: Implementation of simple CNN in latex and python by hand, in scalar and matrix forms. Both training and inference.
- ML: Basically, a lot of ML algs, including SVM, decision trees, random forests, boosting, clustering, PCA, t-SNE, etc.
- perf: Cuda, speeding python up with mutltithreading Rust, SIMD in Rust etc.
- vis: Texty.org-like visualizations using Altair.
- sem_3
- academic_writing: Writing a paper on the topic of the course.
- camera_calibration: My thesis project.
- CV: Computer vision course. All kinds of stuff, including classical, deep learning. Segmentation, detection, tracking, etc.
- health: A project with Toronto University.
- MMDS: Mining massive datasets. Spark, Streams, BloomFilter, etc.
- RL: Reinforcement learning. Q-learning, DQN, A3C, etc.