Inspiration
Delayed and erroneous diagnosis by medical professionals account for nearly 800,000 deaths or permanent disability in the USA every year. We believe we can help cut down this number.
During the hackathon, we also ran into collaboration issues as members wrote code that we each could not sometimes figure out. We believed that many teams face such issues and thus we created a solution to this. ParallaxAI can also therefore solve this issue
What it does
ParallaxAI uses the gpt3.5-turbo model. It collects any large piece of data(e.g long research article about a rare disease) and generates a summary of that article of text. This cuts down the time that medical professionals have to spend reading long pieces of articles.
How we built it
We used React for the frontend, and express for the backend.
Challenges we ran into
- OpenAI's API keys provided to us by the organizers sometimes got revoked and would set us back in testing our project's functionality
- Git conflicts occured more often than we thought. Although we were working on different branches, git merges had to occur on the main branch at some point
Accomplishments that we're proud of
- ParallaxAI works
- We set up a Github workflow to prevent api key revokes
What we learned
- collaboration
- Github workflows
- Working with the gpt3.5-turbo
What's next for ParallaxAI
- Text to audio for summary
- Mobile application
- Working on the scalability of the app
- Making the UI more beautiful
Built With
- express.js
- openai
- react
Log in or sign up for Devpost to join the conversation.