A few works, I made and published.
Python • LangGraph • ChromaDB • OpenAI • API
A customizable RAG agent built with LangGraph, ChromaDB, and OpenAI models, featuring a Telegram Bot frontend. It indexes PDF files, stores embeddings, retrieves relevant chunks, and answers queries conversationally. Fully modular, template-driven, and easy to extend.
Docker • Streamlit • GCP • LightGBM • SQL
Machine learning project predicting home credit default risk using the Home Credit dataset. Includes EDA, feature engineering, model training (LightGBM, XGBoost, ensembles), handling class imbalance, hyperparameter tuning, and deployment via Streamlit and Docker.
Python • Flask • SHAP
Machine learning project predicting stroke risk using patient demographic, health, and lifestyle data. Includes data cleaning, feature analysis and a Flask app for inference. Key insight: stroke risk increases with age, with BMI and glucose levels as important factors.
Python • Sci-kit learn • SVM • KNN • Random Forest
Predictive analysis of customer interest in a COVID travel insurance package. Uses ensemble ML models (SVM, KNN, Random Forest) on demographic and travel data to identify key factors, with Annual Income being the most significant predictor.
Biochemistry • Scientific writing • fatty acid
I wrote a popular science article explaining the importance of essential unsaturated fatty acids in human health, their dietary necessity The article translates biochemical concepts into accessible science writing.
stem cells • in vitro cell culture • chondrogenesis
Conference report from the XXXIV Conference on Embryology (ACTA BIOLOGICA CRACOVIENSIA). Title: Chondrogenic differentiation of human mesenchymal stromal cells derived from adipose tissue