Ivan Shepitko Email & Phone Number

Moscow, Moscow City, Russia

• Development of statistical Rating model for the Largest corporate clients of Gazprombank (residents and non-residents of Russian Federation). Solution is based on shadow ratings of rating agencies and shows high performance. The main result is reducing cost of risk (provisions) more than 1.5 times. Performance of models are high according to common benchmark (AUC exceeds 0.75, compared to baseline of 0.51). The area of application – provisions estimation, risk-based pricing and RWA (risk-weighted assets) calculation: for internal purposes and to meet the regulatory restrictions.• Development and validation of statistical Rating model for assessing the probability of default (PD) of Large corporate groups which includes at least one borrower of Gazprombank. The main result is reducing cost of Risk more than 1.6 times. Performance of models are high according to common benchmark (AUC exceeds 0.75). Additionally, development of Group support module for adjustments PD-estimation for companies provided with significant connection with parent group. This adjustment increases efficiency of final rating model more than 5% (relative). The area of application – provisions estimation, RWA (risk-weighted assets) calculation.• Development and validation of statistical Rating model for Commodity traders. Solution is applicable for the whole Gazprombank group (GPB-group), i.e., for the portfolios of GPB-Russia, GPB-Luxembourg, and GPB-Switzerland. The main result is reducing Cost of risk more than 30% (relative). The area of application – provisions estimation, risk-based pricing and RWA (risk-weighted assets) calculation.• Development and implementation of the methodology for assigning ESG ratings to corporates – clients of GPB-Russia. Newly developed solution for smart and responsible investments.

Moscow, Russian Federation

• Development and validation of statistical models for assessing the probability of default (PD) of Corporate/Commercial borrowers and Small/Medium Enterprises borrowers of DenizBank (top-5 bank of Turkey). Additional point is development of warning signals module for detecting rare events (unusable in statistical analysis) affected PD drastically, such solution enhances level of conservatism (i.e. keeps risk appetite under stronger control) and increases efficiency of final models more than 5% (relative).• Development and validation of statistical models for credit risk estimation (Probability of Default) for Holding Companies and Leasing Companies (two different business segments) – residents of Russian Federation. Additional point is development of Group and Government support module for assessing possibility of group support for companies – members of market group, such solution increases efficiency of final models more than 10% (relative).• Development of statistical model (ML-based solution) for operational risk estimation for Corporate clients of Sberbank. Solution is applied for reducing costs in case of proceedings.

Moscow, Russian Federation

• Development and validation of statistical model for credit risk estimation (Probability of Default) of PIL applicants. Model was developed according to Basel Committee of Banking Supervision recommendations. For more conservative estimation of inflow customers Reject Inference technique was used, it allows to increase final efficiency more than 10% (relative).• Development of scoring system (ML-based solution) for detecting internal and external fraud cases. The main feature of such solution is month-to-month score-track reflects possibility of fraud in near future for target employees.• Development of statistical model for Collection Strategy. During development process was shown that early warnings for customers with delays in repayment (non-default) reduce total loss significantly.• Quarterly monitoring (and reports preparation) of all models for assessing portfolio (retail) risk.

Master Of Science - Ms, Computational And Applied Mathematics, 4.5

Dissertation: "Mathematical model of Russian Banking sector"

Bachelor'S Degree, 4.7