cover User

At T-Bank, I owned analytics for 10+ A/B tests and quasi-experiments across credit and insurance products serving 35M+ customers, identifying US$150K+ in projected monthly value.

I deployed a CatBoost propensity model that increased utilization conversion by 1.15 percentage points and an A/B-tested speech-to-text NLP pipeline that lifted offer-to-application conversion by 1.5 percentage points. I also redesigned a performance-incentive system that saved US$270K per month and reduced cost per utilization by 14%.

At X5 Tech, I led causal measurement across a US$100M+ portfolio, using doubly robust, synthetic-control, and panel-data methods to estimate incremental revenue and adoption. I turn ambiguous growth questions into KPIs, evaluation designs, robustness checks, and clear rollout recommendations.

My research and teaching work combines econometrics, machine learning, Python, NLP, and reproducible data pipelines. I’ve built analytical panels from millions of records, developed multimodal ML features, and led seminars in Python, ML, econometrics, NLP, and recommender systems.