01
The Problem
Information is siloed across different file types including text documents, images, and PDFs, with no unified way to search and retrieve across all of them semantically.
02
What It Does
A RAG pipeline that indexes and retrieves across text, images, and PDFs using GPT-4o's vision capabilities combined with FAISS for fast semantic search and accurate cross-modal retrieval.
03
Why I Built This
I wanted to explore the frontier of multimodal retrieval and what GPT-4o could do beyond chat. Unified document search felt like an immediately valuable use case to test the limits of the technology.