Василь

Rusyn Vasyl
[открыть контакты](см. выше в блоке «контактная информация») | [открыть контакты](см. выше в блоке «контактная информация») | https://github.com/VasyaR

As a Person
I love going to the gym (working out since September 2019) and enjoy other sports, especially competitive ones.

I am passionate about Japanese language and culture, and I achieved the N3 level in Japanese (equivalent to
English B1) through self-study.

I am deeply interested in self-improvement and strive to better myself every day.

I am not a junky or an alcoholic, and I have no criminal record yet. :)

I have a cat and love animals in general.

Education
Lviv Polytechnic National University Lviv, Ukraine
Bachelor’s degree in Computer Science (AI specialization) Sept. 2021 - June 2025

Projects
Bachelor work | Python, Pytorch, Gradio, Huggingface, Kaggle September 2024 – June 2025
• Developed a SOTA kanji recognition model using CASIA-HWDB and Kuzushiji-kanji datasets.

• Used Preact ResNet-18 as a backbone with pytorch and my own concept of cross-language transfer learning

• On these results I’ve published an article in the MDPI journal (Q1 in Web of Science and Q2 corresponding to

Scimago Journal Rank) https://www.mdpi.com/2076-3417/15/9/4894
• Developed additional model for Confidence thresholding for the unkown kanji

• Created a web application using Gradio and deployed it on Huggingface

https://huggingface.co/spaces/Beav3r/KKanjiRecognizer
• Used Kaggle Env for training

• Created a python package https://github.com/VasyaR/KKanjiRecognizer

Dark humor generator | Python, Gradio, Huggingface November 2024 – Dcember 2024
• Developed a RAG system as a chatbot that generates dark humor jokes using LLMs without any censorship (so it

can be offensive and users should be careful)
• Created UI using Gradio and deployed it on Huggingface

https://huggingface.co/spaces/Beav3r/Dark humor generator
• Used Groq llama as a base model

• used bm25 and semantic embeddings from distilroberta to find the most relevant docs (User can regulate the

impact of each)
Studrate | Python, Flask, React, PostgreSQL, Alembic, Nodejs September 2022 – June 2023
• Developed a full-stack web application using Flask serving a REST API with React as the frontend; Also nodejs

web-sockets for web-site‘s chat
• Implemented OAuth

• PostgreSQL as database with alembic for migrations

Burning Revenge | Unity, C# October 2021
• Developed a 2d shooter-platformer using Unity Engine

• Uses custom animations for characters

• Was created during a hackathon with my groupmates
Languages
English B2, Japanese N3, Ukrainian fluent

Achievements

• Scientific article in MDPI journal (Q1 in Web of Science and Q2 corresponding to Scimago Journal Rank)
https://www.mdpi.com/2076-3417/15/9/4894
• ACM ICPC 2022 1st stage: 256/762
• Scored Rating 90.6+/100 for each term of studying in Lviv Polytechnic National University CS AI what is 3-13 out
of 326 students
• Epam Business Analysis course 2023
• Udemy: Data Warehouse course - The Ultimate Guide by Nikolai Schuler
• EDX: Hello (Real) World with ROS – Robot Operating System

Technical Skills
Languages: Python, C/C++, SQL (Postgres), JavaScript (React), HTML/CSS, C#, scala
Frameworks: React, Node.js, Flask, PyTest, PyTorch, CUDA, OpenMP, OpenMPI, pyenv, Gradio, LLM usage, ROS,
Unreal Engine 4
Developer Tools: Git, Docker, VS Code, Visual Studio, IntelliJ, Swagger
Interesting areas of work: GPUs programming, RL, NLP, GameDev, CV, other AI/ML connected