Inspiration

We both started grad school not long ago and after the recent rise of LLMs in the world we got to thinking about how language models could improve the learning experience of students.

What it does

The project allows students to use course materials(pdfs, video transcripts, group conversation, ppts) as context for their personalized LLM Agent.

How we built it

We built it by using a number technologies. Python, Javascript, Django, LangChain, Canvas API, OpenAI API, Vector Store(In memory, FAISS), HTML, CSS

Challenges we ran into

  • storing context for each conversation
  • Build Files indexes and Vector Store
  • Issues with display content on pages

Accomplishments that we're proud of

We are proud to showcase a prototype of a tool that could change the way students learn and also proud of all we have learnt to build it.

What we learned

  • LLMs are expensive and quite difficult to deploy and run in production

What's next for AI Study Buddy

  • Current we are only parsing pdf content, but we look forward to building all other integration necessary to bring the full vision to light.
  • Provide more control over how users use indexed files as context

Built With

Share this project:

Updates

Submission history