SketchyAI

TrainingSet

Inspiration

We wanted to build something that could relieve the pain of building applications for beginners and more experienced users. This could potentially help those with no coding experience build apps and those with coding experience would be able to save time in development.

What it does

The piece of software is able to turn a simple UI drawing into an entire mobile application.

How we built it

We used react.js to build the drawing environment for our users. We trained our images using yolo (a machine learning CNN algorithm) in google cloud compute engine. We then used flutter to be able to transform the classifications from the machine learning algorithm into real code. Finally, we put everything together using flask and firebase to communicate across all these platforms.

Challenges we ran into

We stumbled a lot due to the fact we had to learn many new technologies along the way. Furthermore, time constraints made us have to prioritize certain features.

Accomplishments that we're proud of

We are overall proud of having used many technology stacks and having built something we believe is really cool. We are also very proud of how much we learned along the way.

What we learned

We learned how to deploy machine learning models. We also learned a couple of new frameworks like react and flask. The experience with multiple technology stacks gave us an overall sense of how one would deploy an actual application into production.

What's next for SketchyAI

We would like to add multiple features in the near future. We would like to add more color options, the options to represent data logic within a drawing, to save drawings within a workspace, and even have a sharing feature.