Никита

Nikita Sobur
Kharkiv, Ukraine — [open contact info](look above in the "contact info" section) — [open contact info](look above in the "contact info" section) — [open contact info](look above in the "contact info" section) —
My Telegram — Ukrainian

EDUCATION

V. N. Karazin Kharkiv National University,
Kharkiv, Ukraine Enrolled: September 2019 — Graduated: June 2023
Bachelor of Science: Applied Mathematics
V. N. Karazin Kharkiv National University,
Kharkiv, Ukraine Enrolled: September 2023 — Graduated: December 2024
Master of Science: Applied Mathematics

EXPERIENCE

SoftInWay Burlington, Massachusetts, USA, remote job
Machine Learning Engineer, full-time May 2024 - Present time
• I developed ML models using Tensorflow for predicting compressor performance by its geometric design tabular data
(huge dataset 1M+ rows). Implementation of modern models (TabM) from scratch from recent research papers.
• Experience working with highly unbalanced data: class imbalance, skewed long-tailed distributions of continuous
features.
• I applied combination of the differential evolution optimization algorithm (for searching candidate solution) and
Autoencoder neural net (for search space dimensionality reduction) and decreased prediction time by x3.
• I designed custom loss functions to increase the accuracy of the DE optimization algorithm predictions.
• LinkedIn Assistant application designed and developed for company salespeople with data collection on JavaScript,
backend on Flask and integration of Salesforce CRM API. This development increased salesperson’s speed of work
by 40%.
• Designed and developed RAG-based ChatBot Peter with reranker model over company’s document database. It was
used commercial OpenAI LLM API and Milvus Vector DB for storing text chunks and vectors. LLMOps support
and improvement for the built system.
Freelance/Part time/Research projects Kharkiv, Ukraine
Machine Learning/Computer Vision February 2022 - May 2024
• I developed Computer Vision projects for solving Semantic Segmentation (it was used U-Net model for segmentation
of tumor images), Object Detection (with YOLO models family), Classification and Image Generation tasks.

PROJECTS

Physical models Github repository Kharkiv, Ukraine
Project Link: https://github.com/nikitosssobur/Physical-models
July 2023
• I came up with the idea of creating general repository, which contains different physical models. These models are
developed using mostly Pymunk and Pygame libraries with Object oriented approach.
My-ML-Projects Kharkiv, Ukraine
Project Link: https://github.com/nikitosssobur/My-ML-Projects
November 2023
• I created repository with my ML projects. Here I can demonstrate my skills of using ML libraries such as Pytorch,
Sklearn, Pandas and implementations of models for solving different tasks.
GANs-and-Roses Kharkiv, Ukraine
Project Link: https://github.com/nikitosssobur/GANs-and-Roses/tree/main
November 2023
• My Pytorch implementations of the different types of GAN (Generative Adversarial Networks) models.
Rope-Model Kharkiv, Ukraine
Project Link: https://github.com/nikitosssobur/Rope-model
March 2025 - May 2025
• This is my research project in the intersection of mechanics and numerical modelling of the physical processes.
A rope system similar to those used in rope jumping was considered. The project was developed from scratch,
differential equations were derived and programmed to find a numerical solution using the SciPy library and the
Python programming language. It is builded open-source CLI executable program for experiments.
SKILLS

• Programming: Python, JavaScript, SQL.
• Libraries and frameworks:
Machine Learning and Scientific computing: Pytorch, Tensorflow, Sklearn, XGBoost, Numpy, Matplotlib,
Scipy, Sympy, Pandas, Seaborn, Pymunk, Pygame, Optuna, Keras-Tuner.
Computer Vision: OpenCV, torch-vision, PIL.
LLM with RAG development: Milvus, FAISS, LlamaIndex, transformers, sentencetransformers, Ray.
Backend development: Django, FastAPI, requests, gRPC, PostgreSQL, REST API, JSON.
Data parsing and validation: Selenium, BeautifulSoup, Pydantic.
• Software: VS Code, PyCharm, Jupyter Notebook, Google Colab, Github, Gitlab, Git, Teams.
• Communication: Ukrainian (Native speaker), Russian (Native speaker), English (Intermediate).
• Soft skills: Tutoring/Mentoring, Critical Thinking, Attention to details, Searching information, Asking questions,
Problem Solving, Research, Ideas generation.

INTERESTS

Mostly, I’m interested in machine learning and deep learning fields. This is a huge and fast-growing area with exciting
development opportunities.
Desired positions: Machine Learning Engineer, LLM Engineer, Machine Learning Researcher, Data Scientist, Python
Backend Developer, Computer Vision Engineer.