Inspiration
Inspired by the bad posture of all our team members, we decided to create a tool to help people who work at their desks sit .strAIght. In addition, bad posture is not just some benign problem. A 2024 study found that medical costs and absenteeism due to posture-related musculoskeletal issues cost between $45-54 billion annually.
What it does
Our website will use your webcam to detect how good your posture is using LLM calls, and even provides scores, profile history, notifications, and recommended exercises for you to improve your posture.
How we built it
We used both flask and Boundary ML's BAML in our tech stack. BAML is a programming language that allows easier interfacing with LLM models for computer vision. The site takes periodic photos using the webcam and utilizes BAML to send requests to the LLM. BAML lets us guarantee a return in the form of a description and a score 1-10 based on the user's posture. The photos are saved in the profile history, and the descriptions are compiled and also used to give the user exercise recommendations.
Challenges we ran into
Midway through our project, we needed to rewrite our tech stack as we were facing too many problems writing it in javascript and migrated to flask.
Accomplishments that we're proud of
We are proud that even though this is the first hackathon for all our members, we were able to have a deliverable product that is fully functional while also learning tech we have never worked with before.
What we learned
Through building this project we all familiarized ourselves with the design process, collaboration and also specific technologies like flask and BAML.
Built With
- baml
- boundaryml
- css
- flask
- html
- javascript
- python
Log in or sign up for Devpost to join the conversation.