Metamorphosis (Team OREO)

Inspiration

We wanted to create an AI assistant for the changemaking process that provides common-sense responses to the users, backed by scientific evidence.

What it does

It utilises Retrieval Augmented Generation and Web Search tooling, together with a Large Language Model to provide a chatbot interface for users.

How we built it

We built a streamlined workflow as follows: Data Analysis, Retrieval Augmented Generation, Web Search, LLM Output.

We built a workflow that went through the above stages before finally giving the user a response to their question regarding change-making.

Challenges we ran into

Initially, we only had a pure RAG component, which was very inflexible. We iterated on this to

Accomplishments that we're proud of

We managed to get tangible results from our data analysis, that improved our app efficiency by almost 10 times.

What we learned

As we were unfamiliar with AI, this was a good experience for us to learn about the different tooling available.

What's next for Metamorphosis (Team OREO)

Multi-user support

Built With

• langchain
• python
• streamlit