Automatization of credit scoring pipeline from 20% to 90% level. 4 times growth of Revenue without raising additional underwriting costs.Improve current credit pipeline and scoring models, portfolio management approach. Expand data science and data analytics team from 4 to 10 people.Development of credit scoring models, monitoring, and analytics tools:- Collect data from distributed services and clarify the decision-making process- Develop BI dashboards with key credit metrics, vintage analysis- Highlight weaknesses of current process and models and fix it- All current models are rebuilt with GINI growth up to 10 pct.- Develop new credit limit algorithms based on model- AlertingStack: Python (lightgbm, catboost, flask, sqlalchemy, scikit-learn, pandas), Postgresql, SQL Server, Power BI, Git, Grafana

BI application development for the medical field:- capacity utilization by branch, doctors, or type of services- CJM and internal funnel- Finance&managment accounting: revenue and costs allocations, calculate financial metrics, report structureI took part in all steps from raw data to the final view (raw data investigation, ETL, data marts, visualization)Ad-hoc analysis:Analytics of customers-branches geo and modeling new branch performance depending on location, through the following steps:1) Geocoding of client's addresses, branches, concurrents locations2) Add additional geoinformation: equity prices, infrastructure, etc.3) Data preparation: cleaning, resampling4) Building and validation of the model5) Apply the model to geodata and automatically choose the best locations6) Provide some additional maps and visualizationsStack: Power BI, SQL Server, SQL profiler, python (pandas, sklearn, numpy, plt, plotly, folium, geometry, geopandas), git