Michael Sokolov Email & Phone Number

Russia

Moscow, Russia

◦ Leading the recommendation system development for shop.mts.ru.◦ Built and launched complementary recommendations using prod2vec and associative rules approachresulting in 25% add2cart rate increase and boosting item coverage from 34% to 89% .◦ Taking part in development of RecSys platform and the open-source recommendation systems framework RecTools.◦ Mentoring junior developers.

Moscow, Russia

◦ Rebuilt end-to-end churn prediction ML pipeline which led to 50%-reduction of churn rate and increased YoY money effect by 93%.◦ Developed customer acquisition ML model based on geoanalytics features increased conversion rate by 80%.◦ Implemented MLOps approach using MLFlow, AirFlow and GitLab within the team which standardized the continuous delivery of models in production.

St Petersburg, St Petersburg City, Russia

◦ Used regression models to predict the most efficient amount of products to be sent to stores in order to increase business margin.◦ Automated the process of minor problems solving by implementing clustering method which decreased average SLA by 40%.

◦ ML-engineering: Applying ML-methods for attributes and entities classification which helps in structuring business processes.◦ Analytics: Metadata analysis. Setup of conceptual and logical business models. Developing the pipeline of Information Systems analysis.