Work
Work Experience
Internships and research roles where I turned open-ended problems into working systems.IIT Madras, CAMS IIT-M Fintech Innovation Lab
June 2024 - August 2024Research Intern · Chennai, Tamil Nadu
Explored how large language models (LLMs) can transform financial analytics. Fine-tuned a Retrieval-Augmented Generation (RAG) pipeline on domain-specific datasets to improve data-driven modeling and automated report generation.
- Conducted feasibility analysis across multiple AI-driven financial models to assess LLM integration in fintech workflows
- Fine-tuned a RAG-based LLM on financial datasets with web-search capabilities, improving retrieval precision by 30%
- Evaluated the system's precision on real-world fintech use cases, deepening understanding of AI in risk modeling, compliance checks, and financial decision systems
AI/MLLLMsRAGFintechPython
Renault Nissan
September 2024 - November 2024Technical Intern · Chennai, Tamil Nadu
Designed a data intelligence system to streamline vendor selection and supply chain decisions as part of the Renault-Nissan Industry Innovation Lab.
- Built an algorithm that analyzed 200+ supplier histories—including price consistency, defect rates, and delivery timelines—to recommend the most reliable partners
- Engineered vendor evaluation system reducing decision-making time by 20%
- Created a chatbot interface to provide instant access to vendor insights, cutting manual lookup time by 40%
- Demonstrated how AI-driven analytics can optimize manufacturing operations in large-scale automotive ecosystems
Data AnalyticsAISupply ChainChatbotPython
FlowNow
March 2025 - June 2025Technical Intern · Berkeley, California
Developed an agentic AI system for a neuroscience-based edtech startup from UC Berkeley's SCET network that personalizes reading comprehension through adaptive question generation and focus tracking.
- Designed an agentic AI system generating context-aware questions from user-submitted books to assess comprehension
- Built model analyzing user attention span using behavioral metrics and dynamically adjusted task complexity
- Modeled user attention spans via behavioral metrics, delivering real-time feedback and engagement analytics
- Integrated a Tawk.to-based support agent to improve onboarding efficiency, reducing response times by 35%
- Combined interests in cognitive AI, product design, and user psychology in building AI-driven learning systems
AIEdTechNLPUser AnalyticsProduct Design