Generative AI Hackathon with IBM Granite

Start

CogniverseAI team excels in multi-agent, multi-model systems, specializing in complex LLM pipelines.

CogniverseAI

uvtechnologyins

Uday Valera

AI-Powered Interactive Roadmap Generation for Learning

Our system transforms unstructured documents into structured, interactive learning roadmaps using Retrieval-Augmented Generation (RAG) and Chroma as a vector store. It helps learners navigate complex materials efficiently by consolidating key topics into logical, visually structured nodes.

How It Works:
Chunking & Metadata Extraction – Documents are split into chunks (~1000 tokens), each tagged with chunk_id, topic, keywords, and summary for efficient retrieval.
Roadmap Generation – Instead of mapping chunks directly, the LLM organizes related content into structured learning nodes, ensuring a logical study flow. Each node is assigned a difficulty level to help learners prioritize topics effectively.
Quiz-Based Reinforcement – AI-generated quizzes for each node encourage active learning. Questions reference chunk IDs to maintain accuracy and relevance.
Interactive Q&A – Users can ask topic-specific questions, and the system retrieves relevant chunks to generate informed responses. If gaps exist, it suggests further reading.
Web-Based Visualization – The structured roadmap is displayed in an interactive web interface, allowing users to explore topics, track progress, attempt quizzes, and engage with AI-powered insights.
This system enhances comprehension, retention, and engagement by transforming static documents into an intuitive, adaptive, and interactive learning experience.

An AI-powered system that transforms documents into interactive learning roadmaps with quizzes, chat interactions, and personalized study paths using RAG and Chroma.

"CogniverseAI team on Generative AI Hackathon with IBM Granite Hackathon"

Team Idea

Submission