• File

Ярослав

Data scientist

Considering positions:
Data scientist, Digital-маркетолог, Python-програміст 
Age:
19 years
City:
Lutsk

Contact information

The job seeker has provided: Phone numberEmailMessenger

Name, contacts and photo are only available to registered employers. To access the candidates' personal information, log in as an employer or sign up.

Uploaded file

Quick view version

This resume is posted as a file. The quick view option may be worse than the original resume.

Yaroslav Stetsiuk
Junior Machine Learning Engineer / Data Scientist
Rivne, Ukraine • [open contact info](look above in the "contact info" section) • [open contact info](look above in the "contact info" section)

PROFILE

Junior ML Engineer / Data Scientist, 4th-year student at NTU “KhPI” (Applied Mathematics, program “Intelligent
Data Analysis”). Hands-on experience with SQL and Python at a fintech company and a portfolio of 6 ML projects:
Tabular Data ML (regression, classification, clustering, feature engineering) with Python / scikit-learn, neural
networks and Computer Vision with PyTorch (custom training loops, Dataset/DataLoader, augmentation, GPU).
My workflow: EDA → feature engineering → model → validation (RMSE, R², Accuracy, F1, cross-validation) →
business insights; I control overfitting and data leakage (fitting transformers on train only, sklearn Pipeline). I
understand algorithms “under the hood” thanks to my mathematical background — from gradient descent to PCA.
Seeking a Strong Junior ML Engineer position to work on improving the accuracy of CV/NLP models and
developing new ones.

SKILLS

Python Pandas • NumPy • Matplotlib • Seaborn • ML Pipelines • Clean Structured Code

Machine Learning Scikit-Learn • Linear Regression • Polynomial Regression • Logistic Regression •
Decision Tree • Random Forest • SVM • K-Means • DBSCAN • Mean Shift • PCA •
Pipeline • StandardScaler • Feature Engineering • RMSE • R² • Accuracy • F1 •
Tabular Data ML

Deep Learning / nn.Module • Custom Training Loops • Dataset/DataLoader • BatchNorm •
PyTorch Augmentation • Adam • Image Regression • Image Classification

SQL CTE • Window Functions • JOINs • Subqueries • CASE • Aggregations • Data
Preparation for Analysis

Mathematics Linear Algebra • Calculus • Discrete Mathematics

Analytics EDA • Tableau • Excel • Pivot Tables

Tools Git/GitHub • Jupyter • Google Colab

WORK EXPERIENCE

Present Fintech Company
• Used SQL for data retrieval, aggregation and preparation in a fintech product:
transaction, customer and reporting data.
• Used Python (Pandas, NumPy) for data automation, analysis and processing.
• Automated reports with Python.

PROJECTS

2026 – Present CIFAR-10 Image Classification — Computer Vision (PyTorch)
• Trained a neural network (MLP, ≈1.05M parameters) to classify 60,000 CIFAR-10
images into 10 classes.
• Built a pipeline: Dataset/DataLoader, augmentation (RandomHorizontalFlip,
RandomCrop), BatchNorm1d, CrossEntropyLoss, Adam; GPU (CUDA) training.
• Wrote custom training and validation loops (loss/accuracy), weight saving (.pth) and
prediction visualization.
• Technologies: Python, PyTorch, torchvision, Matplotlib, Google Colab.

2026 – Present Medical Cost Prediction — Tabular Data, Neural Network Regressor
(PyTorch)
• Built a neural network (nn.Module, 10→100→50→1, ReLU) to predict medical costs
from age, BMI, smoking status and region (1,338 records).
• Feature engineering: one-hot encoding of categorical features; identified smoking as the
main cost driver (correlation ≈ 0.79); standardized X and y with fit on train only (no
data leakage).
• Custom training loop (MSELoss, Adam, 600 epochs); evaluation in real units via
inverse_transform (RMSE, R²).
• Technologies: Python, PyTorch, Pandas, scikit-learn, Seaborn.

2026 – Present House Price Prediction — Regression, California Housing (scikit-learn)
• Compared 4 regression models: the best was 2nd-degree polynomial regression, RMSE
65,502 vs 112,967 for simple linear regression (−42%).
• Demonstrated overfitting on the 3rd-degree model (RMSE 69,487) and selected the
optimal model complexity.
• Scaling with fit on train only (no data leakage); PolynomialFeatures + LinearRegression
in an sklearn Pipeline.
• Technologies: Python, Pandas, NumPy, scikit-learn.

2026 – Present Customer Segmentation — Unsupervised Learning (scikit-learn)
• Discovered hidden customer groups by age, income, spending and purchasing behavior
without labels (2,240 customers, 29 features).
• EDA, missing-value handling, feature engineering (Age, Total_Spending,
Total_Children, Customer_Since), standardization.
• Compared K-Means (k = 6, elbow method), Mean Shift and DBSCAN; visualized
clusters via PCA; described the profiles of 6 segments for marketing.
• Technologies: Python, Pandas, NumPy, scikit-learn, Matplotlib, Seaborn.

2025 – Present “BikeStores” Sales Analysis (SQL, Excel, Tableau)
• End-to-end sales analysis: SQL queries for data retrieval and aggregation, Excel pivot
tables, interactive Tableau dashboards.
• Drew business insights on sales dynamics and top-performing product categories for the
final report.
• Technologies: SQL, Excel, Tableau.

EDUCATION

Present Applied Mathematics, National Technical University “Kharkiv Polytechnic
Institute”
4th-year student. Program “Intelligent Data Analysis”. Key courses: calculus, linear
algebra, database organization, discrete mathematics.

LANGUAGES

Ukrainian Native speaker English B1

More resumes of this candidate

Similar candidates

All similar candidates

Candidates at categories

Candidates by city


Compare your requirements and salary with other companies' jobs: