Semantic MCQ Search Engine from a 16,000-Page PDF - RAG / ChromaDB

The client needed a desktop application to quickly retrieve MCQs from large medical textbooks spanning ~16K PDF pages. Keyword search was impractical due to content size and complexity. I designed and built a Python-based semantic search system using a vector database. MCQs were accurately extracted from PDFs using rule-based parsing to handle multiple formats, then converted into embeddings for semantic retrieval. The final app enables instant, highly relevant MCQ search, significantly reducing manual effort and improving study efficiency.