At EPAM Systems, I design and analyze A/B tests of product images and descriptions on brand websites, using click-through rate to retailers as the primary metric. For a nine-brand rollout, I estimated sample sizes for five parallel tests and used the results to recommend which variants to publish.
I analyze monthly active users and 30-day activation rates across partner teams in Power BI. After changes informed by team feedback, adoption of a core platform workflow rose from about 20% to 40–60% of eligible users in a typical team.
I automated data extraction from REST APIs, data processing, and recurring report exports with Python, saving approximately 10 hours of manual reporting per week. I also built Power BI reports for tracking metadata completeness, content coverage by market, and time to publish.
In a Junior Data Scientist, Demand Forecasting role at EPAM Systems, I analyzed product sales series and evaluated LightGBM forecasts. I corrected data leakage in lag-based baselines and built ARIMA and ETS benchmarks; earlier, I published website content in WordPress and ran functional and regression tests on web and mobile apps.

