June 2019 - present: Data Scientist in Sberbank of Russia. My main task is to implement ML algorithms into the Bank's employee management system. Main focus is on the development of models predicting deviations in the work of the Bank's employees or their inefficiency, using process mining. Special emphasis on results visualization for business customers and feature importances interpretation. Projects implemented: managers performance predictive model for business relation Division; model for predicting the dismissal of employees. I also worked on the implementation of Automated Machine Learning module into management system (the module for automatic generation of standard features from "raw" data is developed). Technologies used: scikit-learn, Catboost, lightGBM, H2O, TPOT, XGBoost, SHAP, eli5, SQL (Oracle, PostgreSQL).
December 2018 - June 2019: Data Scientist at Yum Restaurants Russia and CIS (KFC Brand). The main tasks are forecasting sales in Yum restaurants in Russia, model deployment at the MS Azure cloud platform. The implemented project is a model of sales forecasting in the restaurants in Moscow (main stack is H2O). Technologies used: ARIMA\SARIMA, XGBoost, H2O, Spark, MS Azure, SQL (MySQL).
till December 2018: Lead engineer at RSC Energia. Worked on the prediction of radiation onboard the International Space Station. Technologies used: scikit-learn, XGBoost