At Carelon Global Solutions, I architected an AI-powered insurance claim appeal detection service using ML, NLP, Explainable AI, and LLMs to classify AI-generated and human-written appeal letters and flag low-justification claims.
I also worked on an intelligent claims assistance audit tool that processes medical claim documents and uses a RAG pipeline to recommend whether claims should be approved or denied.
At MathCo Pvt. Ltd., I implemented an LLM-powered RAG pipeline for hiring processes and developed topic and sub-topic models to surface themes and sentiment in customer reviews. I also built a laptop demand-forecasting model and an attrition analysis pipeline with automated alerts for model drift.
At MathCo Pvt. Ltd., I designed OLS and Bayesian Marketing Mix Models to support marketing-spend optimization. Earlier, at Tata Elxsi Limited, I built tools for drift detection and production resource monitoring, and at Madras Security Printers Pvt. Ltd. I managed database servers for the Government of Arunachal Pradesh's Ration Card Management System.
