MediNote

MediNote-Cover

Inspiration

We spent hours bouncing around ideas–from a podcast summarizer to a resume builder. However, our ideas seemed to miss the mark on solving a real-world problem. That is until we looked to our world's most important people: our doctors and nurses. We reached out to our connections in the healthcare industry and we finally discovered our problem–how can we make patient documentation more efficient and readable? Read about our solution👇

What it does

To tackle our challenge, we wanted to enable doctors to document their notes through speech and deliver a simple-to-read note for nurses to browse through quickly. Center to our functionality is our keyword extraction feature which allows nurses to search for specific patient details in just a few clicks.

How we built it

Assembly AI
Cohere AI
HTML
Figma
React
Express
Node Js
Sass
API

Challenges we ran into

Pulling API data
Git Flow and merging code
We attempted to create a voice recorder but had trouble sending blob data to the back-end
Being able to understand what was needed in the medical setting, and how the technology we wanted to build would fit in

Accomplishments that we're proud of

We were able to transcribe audio into a transcript
We were able to filter the transcript for filler words as well as filter key words that could be highlighted in a medical note

What we learned

Usages for NLP models
Transcribing audio
Training NLP models using Cohere
Working with blob data

What's next for MediNote

This is not the end for MediNote. We see MediNote becoming integrated with current healthcare systems such as Epic. Through more research and iteration, we can find a better user and market fit for the technology that MediNote provides.