Sep 2020 – Aug 2023 · 2 yrs 11 mos · Bengaluru, Karnataka, India

• Underwriting and Portfolio Management
• Spearheaded the underwriting policy for Rely Credit launch, integrating scorecuts based on income, DTI, application attributes, and credit scores. Achieved a 40% fund rate while maintaining a 5% 60+ DPD.
• Developed underwriting rules for populations lacking bureau data, mitigating risks and declining high-risk segment with 10% 60+ DPD during early policy stages.
• Launched a test to offer higher credit lines to 50% of customers, resulting in a 30% rise in lead-to-loan conversions. Subsequently expanded these credit lines for all customers.
• Developed decision tree utilizing TU scores and attributes to identify and decline high-risk bureau hit customers which significantly reduced portfolio bad rates by 25% and enhanced credit quality.
• Partnered with a third party to create a risk model with Order Express transactional data, tailoring it to their customer base and achieving a 20% reduction in defaults.
• Boosted originations / fundings by 40% by relaxing ineffective criteria while tightening criteria like Rely model scorecut with a higher KS statistic and information value.
• Reduced portfolio losses by 15% by implementing line assignment at application stage instead of pre-qualification. With more bureau attributes at application stage to evaluate risk, it ensured that higher loan amounts were funded to low risk customers
• Database Setup and Risk Reporting
• Integrated multiple data sources to establish databases for application, underwriting, third-party, and loan performance, enhancing our risk analysis capabilities.
• Developed comprehensive reporting system, offering insights on application and loan volumes, credit score distributions, demographics, balances and transactions, revenue and charge-off curves. Empowered stakeholders with valuable insights for informed decision-making and risk assessment.

Jul 2016 – Apr 2018 · 1 yr 9 mos · Delhi Area, India

• Software Innovation and Development
• Managed and optimized APIs for Samsung's TVPlus app serving 10M+ daily requests covering registration, authentication, and payments. Developed 30+ new APIs improving billing and licensing.
• Engineered a recommendation system for Samsung's TVPlus using python TensorFlow. Trained models on purchase history, descriptions, posters, genre features. Launched as a Django service and showcased at Samsung Idea Factory.
• Developed an intuitive monitoring dashboard, providing real-time insights into daily process run times, error notifications, and API testing outcomes.

Oct 2015 – Jun 2016 · 8 mos

• Developed 250+ trading alphas by analyzing market and book data. Simulated and selected high-performing alphas based on returns, fitness, and Sharpe ratio, driving significant revenue.
• Effectively managed this part-time employment with academic pursuits. Ranked in top 5% at the WorldQuant Alpha Competition, demonstrating superior analytical and trading skills.

May 2014 – Jul 2014 · 2 mos

• Developed a de-duplication service for the company's databases, which would not allow a new record to be added if similar records already present. The databases held customer information for their telecom client Vodafone. Important for company in a legal capacity and used in production.

May 2013 – Jul 2013 · 2 mos

• Created modules in VHDL (Xilinx) for improving contrast of video streaming from surveillance cameras, followed by background subtraction for object tracking. Code implemented and programmed in FPGAs.