At University of Maryland, I developed shape-based CNN, autoencoder, and SVM workflows to classify odorant responses, achieving 0.95 balanced accuracy under held-out evaluation. I also built baseline-aware machine learning pipelines for artificial-nose experiments.
As a Teaching Assistant (Agentic AI) at University of Maryland, I guide graduate students building autonomous agents and multi-agent workflows with LangChain, LangGraph, and RAG architectures.
At Hackveda Private Limited, I analyzed transaction records to map regional demand and built Power BI dashboards for sales and inventory monitoring. I also used regression and hypothesis testing to assess price elasticity across discount variations.
My projects include RecruiterRadar, an AI agent for vetting recruiter outreach, and DCFare, a taxicab-fare prediction pipeline that achieved an R² of 0.96. I contributed to a posture recognition and correction system at University of Pune, with findings published at IEEE TEECCON 2024.
