Inspiration
As students, we've all been there — staring at 100 pages of notes the night before an exam, completely overwhelmed. We wanted to build something that makes studying smarter and faster. Instead of scrolling through endless pages trying to find that one thing you need, what if you could just ask a question and get an instant answer? That's the idea behind Study Buddy.
What it does
Study Buddy is an AI-powered study assistant. You upload your lecture notes or study materials as a PDF or TXT file. The app splits them into chunks and stores everything securely in Snowflake's cloud database. Then you can ask any question in the chat interface and Study Buddy retrieves the most relevant parts from Snowflake and uses Groq LLaMA3 AI to give you a clear, helpful answer — like having a personal tutor available 24/7.
How we built it
We built Study Buddy using Python and Streamlit for the frontend, Snowflake as the cloud database to store note chunks and user accounts, and Groq LLaMA3 as the AI engine. We implemented RAG (Retrieval Augmented Generation) which splits notes into chunks, retrieves the most relevant ones from Snowflake, and feeds them to the AI for accurate answers. User authentication is handled securely with SHA-256 password hashing stored directly in Snowflake.
Challenges we ran into
The biggest challenges were connecting Snowflake to Python and making queries work reliably, implementing secure user authentication with hashed passwords, and dealing with API rate limits from multiple AI providers which forced us to switch from Gemini to Groq. Getting the RAG pipeline to consistently retrieve the right chunks was also tricky and required tweaking our search logic.
Accomplishments that we're proud of
We're proud that we built a fully working full stack AI app in one night completely from scratch. The app has a secure login and signup system, a clean colorful dark themed UI, and a working RAG pipeline that retrieves notes and answers questions accurately — all powered by Snowflake as the backbone. Seeing it actually work and answer questions from real uploaded notes was a great moment.
What we learned
We learned how to use Snowflake as a production grade cloud database for an AI application, how to implement RAG (Retrieval Augmented Generation) from scratch, how to build and style a Streamlit web app, and how to handle user authentication securely with password hashing. We also learned a lot about working with different AI APIs and handling rate limits under pressure.
What's next for Study Buddy
Next we plan to add vector embeddings for smarter and more accurate note retrieval, support for more file types like Word documents and PowerPoint slides, automatic quiz generation from uploaded notes, study streak tracking to keep students motivated, and eventually a full mobile app so students can access their AI tutor anywhere, anytime.
Log in or sign up for Devpost to join the conversation.