Solutions for continuous user authentication- based on gait data and other mobile sensors.Failed experiments around graph database- to analyse fraud rings and estimate similarity patterns across them.Estimation of Risk Analysis of candidates based on digital presence/ PF passbook/ other publicly available databases for credit lending solutions.Developed Behavioural Analytics systems to detect fraudster based on the interactions with the loan applying app.Setting up pipelines and running simulations for analytic reports of video-kyc calls to understand future potential blockages.Algorithms to detect tampering/ reusing the same NID card for various onboarding apps which saves the company nearly $2000 for each fraudster caught.

Constructed models that determine the minimum rest time needed to be given to a battery after use before its put back on charging, to extend its battery life. Self Battery diagnosis system to determine battery life and indicate the rider if he would need to get it serviced.Rider profile analysis to understand the trends and patterns in the riding, charging behaviour over thousands of session. Classify the riders based on the similarity and consistency on their riding patterns.Efficiently running automated pipelines on Airflow which generates data and reports rich in visualisation providing valuable real time insights to the business team.

Managed allocation of monitory assets to college fests which had a turnover of more than 10 Million over one year. Identified key insights with very little downside on development of campus life , which smoothened the working of various students with administration. Prepared yearly budget with the extra funds to which both the student and staff can be benefited out of. Developed faster and digitalised systems for reimbursements, rewards and prizes to track each students effort being spent for college activities.

Engineering, Face, documents; Obfuscated, encrypted one of the company's IP Omega, so that it could be safely deployed in clients system without the risk of the technology being stolen. Improved the accuracy of face detector model which serves as a starting point to the pipeline of face match and liveness verification.Developed an algorithm to match and achieve a similarity score between the communication address entered by a customer and the OCR detected address from its legal document.Developed an interactive web page for a client which would pass live clicked images for verification process.

Transporting the real treasure - Fooood😋Out of McDonald's, Pizza Hut, KFC and Zomato- this was the only place I could land such a responsibility.Eventually, I took up a job at UberEats that involved everything from delivering orders, making sure right food was collected and handed to right eater, dealt with hangry customers, restaurant managers, figuring out algos on how delivery partners were given orders and loop holes in the system :PIn short- getting out of my comfort zone was the top priority then until I figured what I wanted in life.