Handstander
Our mission is to empower individuals to challenge themselves. We aimed to create a fitness experience that makes handstands more fun than ever. We focus on rewarding the journey more than the result, allowing people to enjoy competition, consistent improvement, and fulfilment.
We trained the model using Tensorflow and used Alpha pose for video segmentation. Building in public with a community of interested early users was a fulfilling experience. We hosted light competitions weekly for the longest handstand hold, which generated excitement.
A 2-month project to release an alpha version of the app.
- Built-in public assembling a committed Facebook group followers of 127 people
- Gathered over 250 videos of different handstands to train the model
- Created a competitive handstand leaderboard for the engagement experiment pre-alpha release
This was extremely exciting and challenging project. The steep learning curve of Swift development, combined with limited funding, posed major hurdles. Eventually, my co-founder and I pursued other promising opportunities that emerged.
