At Tata Consultancy Services, I engineered LLM guardrails that reduced unsafe responses by 45% and achieved over 92% detection accuracy using embedding-based similarity and transformer classification.
I developed production-grade LLM and GenAI applications with RAG pipelines and multi-agent architectures, using LangChain, LangGraph, and Azure OpenAI. I also designed an AI-agent solution for NFR evaluation that improved processing efficiency by 50% and reduced manual effort.
On my Agentic AI-Based Compliance Automation project, I’m developing a multi-agent system for document ingestion, policy validation, and decision workflows. I’m also building Flask microservices that detect risks such as prompt injection, privacy violations, toxicity, and hallucinations.

