At Schneider Electric, I built production multi-agent LangGraph workflows for HR, Finance, Governance, R&D, and Customer Support, including a natural-language-to-SQL agent for business analysts. I also engineered a reusable RAG chatbot component adopted as the standard across conversational AI initiatives.
I developed RAG pipelines and evaluation-driven release gates, and improved agent reliability with checkpointing, bounded tool calls, retries, and human approvals. On a separate project, I built an agentic AI system for healthcare insurance claim validation with traceable decision workflows.
