Inspiration

Our idea basically revolves around the fact that already existing LLMs don't give suggestions on over-the-counter medications. Gemini for example says " I'm not able to help with that, as I'm only a language model. "

We want to make a RAG using the FDA's drug database (contains information about type, warnings, effects etc.) and make the knowledge base for our model to give suggestions for over-the-counter medicine. Since no prescription is needed to provide such drugs, we want to make a model that gives suggestions based on patient symptoms.

What it does

The LATOM Chatbot is a RAG model built upon LlamaIndex and Together AI to recommend quality Over-The-Counter medicine based on your symptoms.

How we built it

It is an RAG-based model built using the llama3 model and llamaindex.

Challenges we ran into

The way to handle chat history was a little bit complicated and the there was a lot of documentation to go through.

Accomplishments that we're proud of

Our model handles all types of queries but specializes in suggesting OTC medicine.

What we learned

Learned about LLMs and the concept of vector databases and embedding models.

What's next for Over the Counter medication AI Assistant

Dockerizing the application, Cloud hosting it and do more prompt engineering to make the model better.

Built With

  • chainlit
  • llama3
  • llamaindex
Share this project:

Updates