Gen AI · RAG

QnABuilderAI

Upload a PDF and automatically generate well-structured questions of any type and difficulty using retrieval-augmented generation.

Live Demo All Projects

Technical Information

Project Type
RAG ApplicationQuestion GenerationEducational Tool
Tech Stack
PythonFlaskMistral AISentence-TransformersPhiDataFAISSLangChainPyPDFReportLabEmbeddings
Topics / Modules
Document ChunkingPage SelectionTopic ManagementQuestion Types & DifficultyExport
Links
Live Demo

Project Summary

A RAG-powered tool that turns any PDF into a question bank. Configure chunking, choose pages and topics, then generate questions across multiple types and difficulty levels – and export the results.

Skills Demonstrated

Retrieval-Augmented GenerationDocument ChunkingEmbeddingsPrompt DesignPDF ProcessingExport Generation

Solution

Documents are chunked and embedded with Sentence-Transformers into FAISS; a Mistral model generates questions grounded in the retrieved context, with fine control over types, difficulty and topics, exported via ReportLab.

Approach

  1. Upload: Accept a PDF and extract text with PyPDF.
  2. Chunk: Let users configure chunking and select pages to process.
  3. Index: Embed chunks with Sentence-Transformers into FAISS.
  4. Generate: Produce questions by type and difficulty grounded in retrieved context.
  5. Manage: Allow topic management and review of generated questions.
  6. Export: Download results as formatted documents.