Павло

PAVLO BARTMANSKYI
JUNIOR DATA SCIENTIST | MACHINE LEARNING ENGINEER
Warsaw, Poland | [відкрити контакти](див. вище в блоці «контактна інформація») | [відкрити контакти](див. вище в блоці «контактна інформація») | [відкрити контакти](див. вище в блоці «контактна інформація») | github.com/Bartmanskiy |
[відкрити контакти](див. вище в блоці «контактна інформація»)

SUMMARY

Motivated Junior Data Scientist and Machine Learning Engineer with strong theoretical foundations and hands-on experience in
machine learning, deep learning, and data analytics. Demonstrated track record of building predictive models, optimizing neural
networks, creating interactive dashboards with Streamlit, and deploying containerized applications with Docker.

TECHNICAL SKILLS

Languages & Core: Python, SQL
Data Processing: Pandas, NumPy, Matplotlib
Machine Learning: Scikit-learn, Random Forest, GridSearchCV, Feature Engineering
Deep Learning & NLP: TensorFlow, Keras, CNN, VGG16, RNN, LSTM, BiLSTM
Deployment & Tools: Streamlit, Docker, Git, GitHub, Jupyter, VS Code

PROJECT EXPERIENCE

Telecom Customer Churn Prediction
github.com/Bartmanskiy/goit-ds-hw-final_project

Metrics: Accuracy: 94.1% | F1-Score: 94.6% | Tech: Python, Scikit-learn (Random Forest, GridSearchCV), Streamlit, Docker
• Developed an end-to-end classification model to forecast customer churn in telecommunications service data.
• Performed extensive data cleaning, exploratory data analysis (EDA), and feature engineering to identify key churn indicators.
• Optimized Random Forest hyperparameters using GridSearchCV, achieving outstanding metrics (94.1% Accuracy, 94.6% F1-score).
• Designed an interactive web interface using Streamlit and containerized the end-to-end application using Docker for scalable
deployment.

Fashion MNIST Classification
github.com/Bartmanskiy/goit-ds-hw-04/blob/main/Hw10.ipynb

Category: Deep Learning / Computer Vision | Tech: Python, TensorFlow, Keras, CNN, VGG16, Matplotlib
• Constructed and benchmarked custom Convolutional Neural Network (CNN) architectures for multi-class image recognition.
• Leveraged fine-tuned transfer learning using the pre-trained VGG16 model to maximize feature extraction efficiency.
• Analyzed model convergence, loss curves, and confusion matrices to mitigate overfitting through dropout and data augmentation.

IMDB Sentiment Analysis
github.com/Bartmanskiy/goit-ds-hw-04/blob/main/Hw11.ipynb

Category: NLP / Sequence Modeling | Tech: Python, TensorFlow, Keras, RNN, LSTM, Bidirectional LSTM
• Implemented sequence-based Natural Language Processing (NLP) models to classify movie review sentiment.
• Conducted performance comparative analysis across standard RNN, LSTM, and Bidirectional LSTM architectures.
• Preprocessed raw textual data using tokenization, sequence padding, and word embedding layers to capture context-aware long-
term temporal dependencies.

EDUCATION

Data Science & Machine Learning Program 2026
IT School GoIT
M.Sc. Computer Engineering 2024 – 2025
National University of Water & Environmental Engineering (NUWEE)
B.Sc. Computer Engineering 2020 – 2024
Kyiv Aviation Institute (KAI)

LANGUAGES

English: Intermediate | Ukrainian: Native