TVS Credit E.P.I.C 7
Strategy Case Competition
National Finalist
Ranked Top 10/24000+ teams for proposing a comprehensive Gen AI-powered end-to-end lending solution for TVS Credit. The objective was to enhance credit decisioning and customer experience, specifically targeting high-risk, high-growth Tier 2/3 cities and New-to-Credit customers.
Our solution was anchored by a predictive model utilising Multinomial Logistic Regression and Particle Swarm Optimisation for superior risk profiling and underwriting. We addressed the NTC segment by integrating diverse Alternative Data sources (e.g., psychometrics, UPI) and implementing OCR extraction for unconventional financial records like Bahi Khaata.
To transform the customer journey, we deployed Conversational and Multimodal AI bots that offered personalised service and simplified complex loan terms. AI Agents were also incorporated into collections to automate EMI restructuring and deliver targeted nudges. The strategy projected a 7% reduction in delinquency, a 20% cut in operational costs, and delivered a final projected profit of ₹4.26 Cr.
Projects
Impact Compass
Early-stage circular economy startups deliver environmental value but can't attract investment due to fragmented impact metrics. Without standardised data, investors see high risk and corporates struggle to identify credible partners—creating a critical funding gap. EchoImpact solves this by converting operational and environmental outcomes into standardised Impact Scores using 50+ globally aligned KPIs. This enables transparent benchmarking and smarter matchmaking between startups, investors, and corporate buyers.
​
As Business Analyst Intern, I led user research across all stakeholders, translated ecosystem challenges into product requirements, mapped workflows, defined core metrics and scoring methodology, and created platform mockups and the PRD.
The result: reduced information asymmetry, accelerated due diligence, and improved early-stage capital commitments—making circular economy impact measurable, comparable, and investable.
QuizWiz
QuizWiz is a Chrome extension developed to address the challenge of passive online reading and poor information retention among users consuming educational content. Leveraging Generative AI, we built the extension to automatically analyse web articles in real time and generate contextual quizzes based on key concepts. The solution created relevant multiple-choice questions with immediate feedback and explanations, while incorporating adaptive difficulty levels and spaced-repetition algorithms to optimise learning outcomes. This approach successfully boosted reader engagement and improved conceptual retention by 40%.
As a Product Intern, I created the early prototypes, identified potential markets where QuizWiz could be deployed, and assisted in studies to validate our core assumptions around learning effectiveness and user interaction with AI quizzes.
MessMate
As part of our Design Thinking course, my team undertook a study to identify student pain-points on campus, ultimately leading to the creation of a tangible solution. We pinpointed the campus dining experience, marked by hygiene issues, repetitive menus , food wastage and a complete lack of official feedback channels , as a major source of frustration. Our initial approach systematically paired student perspectives with crucial operational constraints from the vendor, such as balancing low costs and managing staff. This dual-sided analysis allowed our team to map where the system was fundamentally failing. We then utilized frameworks like the 5W, Six Thinking Hats, and SCAMPER to collectively distill the root causes. This structured analysis culminated in the development of Messmate, a phased improvement system that introduced weekly hygiene audits , QR-based real-time feedback , and student-voted menu rotations and meal bookings.
