Андрій

Andrii Lysun
Lviv, Ukraine | [відкрити контакти](див. вище в блоці «контактна інформація») | [відкрити контакти](див. вище в блоці «контактна інформація») GitHub: andriyqz | LinkedIn | Telegram: @andriyqz

EDUCATION
Lviv Polytechnic National UniversitySep 2025 — Present
BSc in Artificial Intelligence Systems (2nd Year Student)

TECHNICAL SKILLS
Languages:Python, SQL (PostgreSQL), Git
Machine Learning:XGBoost, PyTorch, Scikit-learn, Feature Engineering, Data Processing
Data Science:Pandas, NumPy, Matplotlib, Seaborn
Backend s DB:FastAPI, Pydantic, SQLAlchemy, PostgreSQL, REST API, Pytest
MLOps s DevOps:Docker, Apache Airflow, DVC, GitHub Actions, CI/CD, AWS
Generative AI & Architecture: LLM, RAG
AI Frameworks & Libraries: LangChain, Hugging Face

PROJECTS
Customer Churn Prediction (End-to-End MLOps)GitHub Repository
Automated churn risk assessment system for E-commerce platforms
• Model Development: Developed and optimized an XGBoost classifier to predict churn probability, categorizing customers into Low, Medium, and High risk levels.
• Backend Architecture: Built a robust FastAPI service with Pydantic validation and SQLAlchemy ORM, featuring a batch prediction endpoint (/predict_batch).
• Data Engineering: Designed a PostgreSQL schema to store customer statistics and historical prediction results, ensuring data integrity.
• Workflow Orchestration: Implemented Apache Airflow DAGs to automate daily batch processing (10,000 records/batch) with built-in fault tolerance and transaction rollbacks.
• Infrastructure: Containerized the entire ecosystem (API, DB, Airflow) using Docker Compose; implemented automated seeding for demonstration environments.
• Testing: Ensured system reliability with Pytest and FastAPI TestClient for integration testing.
Stack: Python 3.13, XGBoost, FastAPI, PostgreSQL, Airflow, Docker, Scikit-learn.

PERSONAL PROFILE
A highly motivated18-year-old developer specializing in Artificial Intelligence Systems.