Part of Fraud Strategy & Analytics team for Marcus Lending PortfolioManaged 3 line of businesses: Personal Loans, Marcus Pay (Buy Now Pay Later) and Line of Credit for Small Businesses• Optimized existing fraud strategies by identifying redundant and low performing rules and onboarded new Fraud Vendors leading tosmoother customer experience (additional ~2M in annual revenue) and reduction in overall friction by 18%.• Automated E2E testing for deploying new Fraud strategies across all the products saving 50 man-hours per month.• Developed automated systems using Python and Tableau to proactively identify unusual spikes across all products and stop fraudulent activities saving $3M in losses per year• Designed KPI reporting framework for Lending portfolio in Tableau which is used by leadership for key decision making.

Part of Fraud Operations Analytics team for Discover Credit Card Products• Developed critical reports that helped track different types of frauds across various segments and devise strategies to prevent them and automate the entire process leading to saving ~$250,000 per month.• Automated the manual categorization of customer complaints and feedbacks, saving 90 man-hours per month.• Streamlined the process of extraction of structured and unstructured Credit Bureau data and loading it into a table in Snowflake using Python for analysts and agents to be able to use it for customer profiling.• Performed A/B Tests on implementation of new fraud strategies and features in a product leading to a reduction in fraud rate by 2.5% and increase in self-service rate by 8%.• Conducted trainings and documented step-by-step process of transitioning to Air9, Airflow and Snowflake leading to quicker and convenient switchover for my team.

Worked as Research Assistant for Department ProfessorsProject 1: Prediction of Ticket Sales for an Ice-Hockey StadiumDescription: • Performed exploratory data analysis and created 20+ user interactive dashboards in Tableau to visualize trends and impact of potential variables affecting ticket sales.• Implemented Regression and Tree models using Python to predict the ticket sales and identify the key features.• Automated the entire pipeline starting from pulling the data from AWS RedShift, running the model to visualizing the key business metrics important for decision making.Project 2: Trends in Medical ReferralsDescription:• Scraped 100K records of doctor details like ratings and reviews from doctor rating websites using Python.• Manipulated and merged huge historical referral datasets (~100 GB) with scraped data in AWS Redshift.• Cleaned the data and performed exploratory data analysis to identify variables that affected the referral patterns.• Leveraged Tableau to forecast the trends and build interactive dashboards and storyboards to convey insights.

Part of Data Engineering Team for a financial services clientRoles and Responsibilities:- Performed ETL (Extract-Transform-Load) activities from different sources to gather information and designed data flows for BI reports.- Optimized performance of a critical set of monthly ETL jobs by reducing its execution time by 30%, improving complex SQL queries used to pull large datasets.- Reduced effort and time took for ETL and various other processes by replacing existing manual systems by Systematic Methodical Automated Process which led to savings of human effort of around 90 hours/month.- Headed a team of five in development and implementation of an efficient process to secure day-to-day loss of transactional data, saving $20,000 per month during a critical period in the project.- Performed root cause analysis to identify problems in the knowledge transfer process, standardized documentation creation, storage, and maintenance process at central repository reducing transition efforts by 50%.- Documented Process flows, Procedure manuals for the functions performed and the training materials.Tools and Technologies: Abinitio(ETL), Oracle SQL, Autosys(Monitoring Tool), MS Excel, Service Now, Tableau