Pratik Priyankar

Fraud Risk Model Validation

• Comprehensive review of true name fraud, synthetic fraud and merchant fraud models.
• Evaluated model assumptions and methodologies for assessment of model accuracy and reliability.
• Conducted quantitative back-testing and out-of-sample testing to assess model accuracy and stability.
• Prepared detailed MRM validation report corresponding to assessment done for the model's accuracy, performance and compliance with regulatory standards. The report details the methodology used, key findings and recommendations for improvements.
• Coordinating with Stakeholders in the US ensuring that the model meets the Model Risk Management (MRM) Internal standards and the regulatory requirements of Model Risk (SR 11-7 by the Federal Reserve, OCC 2011-12).

Model Monitoring

• Monitored various models for changes such as model degradation, data drift and concept drift and ensuring that the models have an acceptable level of performance.
• Calculated KS, GINI, fraud capture rates, PSI, Segmentation distribution and other metrics of various models.

• Working across teams to define problem statements, collect data, build analytical models and make recommendations.
• Involved in Data pre processing Techniques for making the data useful for creating Machine Learning models.
• Developed ML models with round of consultations from various verticals and discussions with stake holders.
• Presenting the model conclusions and insights through various metrics, internal KPIs and graphs.
• Documentation of all the processes and issues occurred in a systematic manner for future reference.

• The project (VM-DATA-OPS) was in the production environment. The data support team ensured smooth transition of data from different sources to the data warehouse.
• Solving tickets and issues related to Jobs failure in SAS9.3 and process flow manager.
• Taking ad-hoc requests by users.
• Performing SAS Admin roles like User Management, SAS Server Smooth running, Archival Process.
• Deployment of new and modified SAS scripts in Production environment.

• Data manipulation by merging, appending and sorting datasets.
• Used PROC SQL to create subsets of data for further analysis.
• Used DATA STEP, PROC MEANS, PROC SQL, MACROS, PROC FREQ, PROC TABULATE, DATA _NULL_, PROC SUMMARY, RETAIN for reporting and analysis.
• Used SAS - output delivery system ODS to format reports in HTML, PDF and CSV reports.
• Optimized SAS code to reduce run time and make the process more efficient.