Mr. Pustovoytov graduated from Moscow Institute of Physics and Technology where he completed stu

Have been developing projects in the field of Data Mining/Data Science/Machine Learning for almost 12 years. Have successfully used various algorithms (including deep learning, NLP approaches and recommending algorithms) to solve problems for various domains including but not limited to financial, retail, FMCG and medicine. Successfully managed projects with teams of up to 50 engineers and researchers.

Last 4 years Mr.Pustovoytov is focused on delivering Data Science-based solutions for financial industry. From 2016 to 2018 he worked on position Head of Data Science in DoubleData, a small Russian startup specialized on developing products and executing projects for financial industry and FMCG brands using publicly available data. Key projects of his team include:

• Predictive propensity model for Top-10 Russian bank with double-digit conversion rate on dormant customers.
• Churn prediction model for Affluent/Mass Affluent segments of Top-10 Russian bank utilizing social network attributes and customer’s data with +30-40% performance compared to best model bank had
• Fraud detection models for several FMCG-brands to clean up CRM
• Predictive models for new Scoring Attribute service, which proved to add 5+ Gini points to bank scoring system according to several pilots (and 10+ for on the hit).
• Computationally effective solution of NP problem for finding a set of places to advertise with maximum coverage considering intersection and budget limitations

Mr. Pustovoytov joined a BCS team in Sep, 2018 with part of former colleagues and established a Data Science department intended to find data-driven solutions for all of BCS businesses. First clients inside BCS include credit marketplace Odobrim.ruand a digital bank for SMB Sfera.ru. In his presentation he will cover several projects accomplished for this businesses.