We initially wanted to experiment with machine learning and see how we can implement it in our everyday life. At the same time, the themes of collaboration, health and wellness interested us a lot. In the end, we thought of the perfect idea that could inspire thousands of people, UYOGA.


UYOGA, a humble place where you can connect to yourself and to others. You can decide if you want to start a Yoga chain of wellness or continue one that had been started by your friends. After a productive yoga session, you can upload a picture of one of your yoga poses as the proof of your workout. If UYOGA validates your yoga pose with proper form, then you can unlock a unique code to continue your chain of wellness. Simply navigate to the friends tab, enter the unique code from the workout you just completed, enter your friend's contact information and UYOGA will send them an invite to continue the chain! From there, your friend can visit UYOGA to start their workout then invite more friends to continue. Doing yoga has never been as fun and collaborative. UYOGA unites the community and brings people together in a unique way. All of this without the hassle of creating an account!

What it does


  • Send a picture of one of the yoga poses from your workout session
  • Get a code to validate your session
  • Enter the code in the 'Friends' tab
  • Refer to a friend to continue your chain!
  • No sign up needed at all

How we built it

The project as a whole was tied together with the flask framework in python. We first made a website prototype on Figma then started writing the HTML and CSS to bring the project to life. A lot of effort went into bringing the machine learning aspect of the project together with the features of the user interface. We also started working on the messaging feature of the website by using Twilio's Messaging API. Last but not least, we implement machine learning, particularly tensorflow in our project for our machine to learn how to identify different yoga poses.

Challenges we ran into

Combining the machine learning aspect with the front end user experience was a challenge. In the midst of it all, it was a bit hard to collaborate on certain files since our team was virtual and it was hard to work on a common workspace. This led to some mishaps on efficiently tying all of the files together as a whole project, however we managed to figure it out through numerous calls and screen sharing late into the night.

Accomplishments that we're proud of

We were quite proud with the final product that we came up with. The way we implemented the machine learning aspect was very unique to our project and with 86.6% accuracy, we are quite proud of what we could accomplish. We also had to work with a lot of new technologies and turns out, there were many things we had to learn on the spot, sometimes it took us hours on end but we are proud that we always managed to find a good solution in the end.

What we learned

Our team had contrasting skillsets; one was more experienced with html and css while the other had more experience with machine learning. We learned a lot from each other and definitely combined our strengths to make the project a reality. We also learned a lot about keeping our code organized, it ended up being a strong asset when collaborating in a team environment.

UYOGA's future developments

One feature that could be extremely fun with UYOGA is to keep track of the longest chains of our users' yoga workouts. See how long a group of friends can keep a chain of wellbeing going. This feature could definitely improve the excitement for the website.

Built With

Share this project: