TRISEVA: Multi-Agent RAG System for Cross-Domain Document QA

Introduction In the current landscape of Large Language Models (LLMs), retrieving accurate information from domain-specific documents is a major challenge. Standard Retrieval-Augmented Generation (RAG) systems often fail when dealing with complex terminology and structured files across highly diverse sectors like Healthcare, Law, and Agriculture. To solve this, I designed TRISEVA (a Multi-Agent Retrieval-Augmented System) as part of my M.Tech dissertation at BITS Pilani. Architecture TRISEVA employs a modular multi-agent architecture where specialized agents collaborate to analyze, retrieve, and formulate explainable answers: