Olga Vopilova
Middle Data Scientist at Mokka- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
-
Русский Native or bilingual proficiency
-
Английский Professional working proficiency
Topline Score
Bio
Experience
-
Mokka
-
United Kingdom
-
Financial Services
-
1 - 100 Employee
-
Middle Data Scientist
-
Oct 2022 - Present
Middle Data Scientist in risk department. - Developed e2e ML solutions - from analyzing new data sources and feature generation to model development and deployment. - I greatly expanded the current feature store in DWH by analyzing new data sources and generating features. - Analysed of the added predictive power of external data sources and calculation of savings from their use. - Developed regular monitoring of current models in order to identify problems with features or metrics. - Implemented PD and behavioral models with a significant improvement in quality compared to current models. - Analyzed four new data sources, generated features based on them, add features to feature store in DWH and developed gradient boosting PD models using new features to improve current model in Russian and European contour. Show less
-
-
-
Sberbank
-
Banking
-
700 & Above Employee
-
Middle Data Scientist
-
Jul 2022 - Oct 2022
Middle Data Scientist in risk modeling department.- Solved problems related to banking risks.- Developed Net Credit Loss model for consumer loans - model for estimating lifetime loan risk at the time of application using gradient booting and linear regression. The model consisted of three drivers (PD model, LGD model, and amortization model) and a final risk calculator that used the outputs of each of the drivers. The model has significantly improved Gini, which has saved millions of dollars for the company.- Conducted weekly monitoring of current models in order to identify problems in the features or metrics of the models- Сonstantly communicated with business customers and the validation department. Show less
-
-
Data Scientist
-
Mar 2021 - Jul 2022
Junior Data Scientist in risk modeling department.- Solved problems related to banking risks.- Developed the Probability of Default model for consumer loans based on a recurrent neural network. The model has significantly improved Gini, which has saved millions of dollars for the company.- Winning the hackathon for solving the task of Take-up modeling.- Сonstantly communicated with business customers and the validation department.- Analyzed the feasibility of using external scores in current models. Show less
-
-
Education
-
Saint Petersburg State University
Bachelor degree, Math and Computer Science