Sai Krishna Tunuguntla

Worked on Unified payment Platform

The platform is a single gateway point for all payments happening in Ola app. The platform acted as payment aggregator for various verticals within Ola like cabs, foods, store, cars.

Was responsible for designing, conceptualizing the platform to provide compatibility with various payment methods like wallets, UPI, netbanking, card.

Successful led the phase one of the project which caters to 1 million payment transactions each day.

Worked on pro-active fraud detection which supplements Ola's legacy rule based fraud detection algorithms.

I was a part of two man team responsible for creation of various anamoly detection models for detecting frauds. We experimented with various machine learning approaches like isolation random forest, DBScan, OneclassSVM and finally made 9 models which were work in tandem to capture about a thousand fraud entities each day.

I was also solely responsible for the feature extraction part which takes customer historical footprints and crunches it into data points for the models to operate on. This complicated pipeline processes about 1.2 Billion events per day.

• Worked on the fraud detection systems at Olacabs. Fraud can happen at any point in a cab booking. Customer, driver, ride, payment instrument, customer/ driver devices. This system is responsible for screening about 2.5 Billion events per day and taking action on fraudsters in real time.
• Was a key contributor in many of the nuanced fraud detection rules and also acted as primary POC for the entire system.
• The project was made by 10 developers in a spam of 6 years and 5 months after i joined the team, almost 6 members left the team Including SDE 2's and 3's. I rose upto the occasion and understood the ins and outs of the system and maintained it for an year before delagating it to others.
• This resulted in me reciving MVP recognition multiple times from olacabs.