AI-Powered Document Intelligence: A Multimoda...

AI-Powered Document Intelligence: A Multimodal Extraction, Semantic Retrieval, and Conversational RAG System for Complex PDFs

Author Name : Deepak Gowda G R, Dr. V. Srinivasan, Dr. Jayanthi R, Bhoomika B R

ABSTRACT This paper presents a production-oriented document intelligence system that unifies multimodal PDF parsing, OCR with image preprocessing, ensemble table extraction, dense semantic indexing with Groq LLaMA embeddings, and a conversational interface for grounded question answering over complex documents. The pipeline integrates PyPDF2/PyMuPDF for text and images,