Looking for a job

Dmytro Bilyk

@dmytrobilyk

Message

Data Scientist at HYS Enterprise who cut e-commerce delivery estimate error by 15% with CatBoost models.

Ukraine

Message

What I'm looking for

I'm interested in ML-focused roles involving financial data, recommendation systems, and forecasting, with responsibilities centered on model development and optimization.

At HYS Enterprise, I developed CatBoost models for customer-facing e-commerce delivery estimates, reducing order-confirmation P50 MAE by 15%. I also calibrated P90 estimates and developed a LightGBM classifier to identify shipments at risk of delay.

At Intelliarts, I built financial data pipelines and trained LightGBM models to forecast ETF portfolio volatility. I helped reduce scenario calculation time from 8.4 to 1.9 seconds, and supported portfolio-allocation calculations using CVXPY.

Experience

Work history, roles, and key accomplishments

Data Scientist

HYS Enterprise

Oct 2025 - Oct 2026 (1 year)

Developed a delivery-prediction system for a regional e-commerce retailer, generating customer-facing delivery estimates and identifying at-risk shipments. Built CatBoost quantile-regression models and a LightGBM delay-risk classifier, reducing order-confirmation P50 MAE by 15% and achieving 74% recall on delay risk.

Python SQL Polars NumPy scikit learn CatBoost LightGBM PyTorch Optuna PostgreSQL AWS RDS S3 Airflow MLFlow fastAPI Docker Grafana

Data Scientist

Intelliarts

Sep 2023 - Aug 2025 (1 year 11 months)

Developed a financial-planning platform that compares investment scenarios and forecasts portfolio risk. Built pipelines for preprocessing and feature generation, trained LightGBM models for volatility prediction, and reduced median calculation time for scenarios from 8.4 to 1.9 seconds.