Особисті дані приховані

PROFILE
Roman Trehub Python developer specializing in AI and machine learning
solutions. Experienced in data processing, deep learning, NLP,
AI/ML Engineer and computer vision, with a strong focus on generative AI, LLMs,
cloud technologies, and scalable data-driven applications.
CONTACT
EDUCATION
Email
[відкрити контакти](див. вище в блоці «контактна інформація») Artificial Intelligence - Master of Science 2025 - Present
(MSc)
Phone
Technical University of Kosice, Faculty of Electrical
[відкрити контакти](див. вище в блоці «контактна інформація») Engineering and Informatics
LinkedIn Currently pursuing MSc in Intelligent Systems with focus on machine
[відкрити контакти](див. вище в блоці «контактна інформація») learning and data-driven methods. Includes NLP, computer vision,
predictive modeling, and development of scalable AI-powered
GitHub systems with real-world applications.
github.com/Romaha095
Artificial Intelligence - Bachelor of Science 2022 - 2025
TECHNOLOGIES
(BSc)
Programming: Python, C/C#, Technical University of Kosice, Faculty of Electrical
JavaScript Engineering and Informatics
Frameworks and Libraries: Flask, Strong foundation in algorithms, data structures, and software
React, FastAPI, Pandas, NumPy, engineering, with focus on data processing, applied machine
SciPy, Matplotlib, Seaborn, OpenCV, learning, and building simple ML pipelines. Worked with databases,
tqdm systems, and networks while developing practical AI-oriented
Machine Learning and AI: PyTorch, solutions and robotics-related projects.
Scikit-learn, Transformers, Hugging
Face, NLTK, PEFT, vLLM, TGI,
LangChain PROJECTS
Databases: PostgreSQL, MongoDB, Search for Mayan Glyph Signs in Dictionaries
SQL Bachelor's Thesis
Cloud and MLOps: AWS, Azure,
Docker, CI/CD, Model Deployment Designed and implemented an end-to-end pipeline to segment
Mayan glyphs from line-drawing transcriptions and match them
Tools: Git, Linux, Jupyter Notebook
against structured glyph dictionaries.
Methodologies: Agile, Scrum
Trained and evaluated a glyph embedding model for
LANGUAGES similarity-based retrieval, enabling automated matching as part of
a larger Mayan text translation workflow.
Slovak - C1 Tech: PyTorch, Transformers, OpenCV, NumPy, pandas
English - A2
Russian - Native Real-time Parking Availability Mobile App
Ukrainian - Native Team project with mentors from FPT Slovakia
Developed backend services for a mobile app displaying available
parking spots with filtering and navigation.
Processed real-time sensor and camera-based data for occupancy
detection and updates.
Tech: Python, Flask, AWS, Docker

Hospital Knowledge Base Search Engine
University Project
Built a search engine for a hospital knowledge base that answers
questions over internal documentation.
Created a document indexing pipeline for PDF files using chunking
and vector embeddings.
Improved retrieval quality with hybrid search and reranking.
Developed a web application that provides answers with links to
the supporting documents.
Tech: Python, LangChain, FAISS, Hugging Face, Ollama, Docker

Roman Trehub | AI/ML Engineer