At Capgemini, I design and ship multi-agent LLM systems for enterprise clients across finance, energy, agribusiness, and procurement. I built a BI modernization platform that converts legacy report definitions into Power BI, Tableau, and Qlik configurations, cutting migration effort by 65%.
I designed a conversational analytics copilot that retrieves context, generates and validates SQL, and produces visualizations. It gave business users self-serve analytics and cut reliance on central reporting teams by 50%.
My work at Capgemini also includes metadata and schema intelligence agents for lineage, business-term inference, and change-impact analysis. These agents removed manual intervention on 70% of upstream breaking changes.
As an AI/ML Intern at Ministry of Electronics & IT (MeitY), I built a real-time sign language recognition system using OpenCV and MediaPipe that reached 95% gesture recognition accuracy. In independent projects, I’ve developed LLM agent orchestration, episodic memory, and requirement intelligence systems.
