Inspiration

During our university time, dealing with formalities and procedures has been a bit challenging. We thought it would be a good idea to help newer generations in this areas.

What it does

It is an AI Assistant, whom the user can ask their doubts related to academic regulations, and provides answers. It can understand written text and speech.

How we built it

For the front-end, we used Flutter, and a Flutter package for speech-to-text conversion. For the back-end, we used Django as a framework; OpenAI as the language and embedding models; InterSystems IRIS for the database, including its Vector Search features for storing and querying the vector embeddings; and LlamaIndex as a data framework.

Challenges we ran into

We had trouble with our lack of experience in this topics and with setting up the InterSystems ecosystem.

Accomplishments that we're proud of

We are very proud of the whole project, since we managed to achieve what we set out to do.

What we learned

We have improved our knowledge in several areas, including front-end, API REST, and working with vector embeddings.

What's next for FIBureaucracy AI

This is just a working prototype, that can be improved and perfectioned.

Built With

Share this project:

Updates