AI Health Assistance

Inspiration

I was inspired by what my community manager said, which goes like this "Be solution-oriented," which pushed me to think deeply about what to build while preparing for a knowledge showcase, which became the output of what I am submitting for this hackathon.

What it does

The AI Health Assistance serves as a rural health Assistance to Nigerians aiming to bridge the gap in healthcare by diagnosing based on the user's symptom/description of what is wrong with them either through text or voice. It also has another feature uploading an image of any affected Area then diagnosis and remedies while at the core of all responses, it references a new Africa-specific medical dataset at the same time emphasising the need to also see a professional if opportunity arises to see one.

How we built it

This simple AI Health Assistant was built using Python, calling open-source models, which are then fine-tuned for the medical use case. Hosted on Streamlit for production.

Challenges we ran into

The challenges I ran into were trying to integrate the Nigerians' realities in terms of relatable voices (text-to-speech), which is still yet to be embedded.

Accomplishments that we're proud of

One accomplishment would be including the Africa-specific med datasets, which are still growing and could revolutionise healthcare access to Africans using LLM.

What we learned

I learned something very important before building. Make sure there is a true purpose behind it because that's what becomes your drive not just building for money's sake as all the world is blindingly chasing after but looking to truly change people in an unexpected way through each build.