First Party Fraud Strategy at Apple Card

Quant- PPNR Model validation, working on high IRR Global Wealth Management statistical and analytical models forecasting UBS group revenues under baseline and stress scenarios including FINMAs LPA as per regulatory guidelines.

Worked on multiple Analytics Projects : 1. Conversion Rate Analysis : Helped understand the consumer psychology while booking a ride on Bounce app and suggested key driving factors that can improve overall conversion rate on Bounce App.2. Demand Supply Mismatch : Bounce has over 1,000 bikes supplied across hexes in Bangalore and Hyderabad area. Came up with a single metric to track DS mismatch over hexes. Huge actionability in terms of improving user experience, bike availability and improving conversion rate.3. Predicting R&M (Repair & Maintenance) incidents using Machine Learning beforehand to reduce R&M Costs and increasing asset life.Participated in Disrupt 2.0 Hackathon at Bounce.

Developed a strong acumen of the required algorithms and techniques and successfully developed AI/DL based modules which holds immense utility in QSR/Retail domain. The modules focus on three use-cases: 1) Audio data based feedback mechanism tool for prediction of age, gender, geography from voice of customer coupled with conversion of the audio to text in real time enabling sentiment analysis and topic attribution. 2) Deep Learning based analysis of final checkout item images using DL based object detection algorithms using Mask RCNN and tally it with the POS records to identify potential mismatch between order placed and order received.3) Analyzing image/video feed of customers in stores by identifying real time mood states from their faces using CNN architecture, thereby enabling improved customer services.