Inspiration
Students often struggle to find specific information from many course PDFs. We wanted to make learning materials easier to search using natural language.
What it does
Users upload course PDFs and ask questions. Nemo AI searches relevant content and generates answers based on the documents.
How we built it
Extracted text from PDFs and split it into chunks. Created embeddings and stored them in ChromaDB. Used semantic search to retrieve relevant information. Used Gemini and LangGraph to generate and manage answers.
Challenges we ran into
Retrieving accurate chunks for user questions. Handling complex questions requiring multiple searches. Removing duplicate and irrelevant results. Accomplishments that we're proud of Built a working RAG-based AI assistant. Implemented semantic search and multi-query retrieval. Made answers grounded in course materials.
What we learned
How RAG systems work. Vector embeddings and ChromaDB. Semantic search, LLMs, Gemini, and LangGraph.
What's next for Nemo AI Assistant
Improve retrieval accuracy. Add more document formats. Add better source references and personalized learning features
Log in or sign up for Devpost to join the conversation.