At Tata Consultancy Services, working with MetLife, I build production-grade agentic AI systems for enterprise insurance workflows. I engineered MCP servers for development and security tools, achieving 95% tool-call success in evaluated workflows.
For claims intelligence, I architected agentic RAG routing between policy documents and claims database tools. The system reached 85% Recall@5 and 90% grounded-answer accuracy, and my evaluation framework measured 93% field-extraction accuracy.
I also built a GraphQL API over PostgreSQL and an agentic pipeline that brings customer submissions and legacy records into an enterprise knowledge base. I supported testing and validation across development, QA, UAT, and production.
Earlier, at Get Better Studios, I built web-scraping pipelines for gym business data and created Power BI dashboards. My projects include a research-paper QA system, LLaMA 3.1 fine-tuning, and fraud-detection models; I also contributed fixes to Unsloth.

