I was hired as a short term contractor to work with the machine learning team to create 15,000+ hand drawn lo-fi wireframe screens of 100+ existing apps available on the Google Play Store. These sketches were used to feed as to a machine learning algorithm to learn different design patterns across different types of mobile applications. The ultimate goal of the team was to create an experimental MVP that would ultimately expedite the wireframing process for designers to make our jobs easier.



Since this project was manually intensive. I knew I needed to expedite the process if I wanted to deliver efficient and accurate results. I needed to create a nimble design system consistent across different apps and different verticals.


Design Strategy

I had to create a design system that expedited the process without affecting the quality of each screen. This design system was super successful as it allowed me to sketch each screen in less than a minute each with a high degree of accuracy.

During the actual sketching and handoff process, I marked each sketch with it's appropriate ID number at the top so when it came time to take screenshots I just needed to cross-check the ID written at the top of each screen during the file naming process. Photos were taken with my mobile phone and sent to myself via text so I could easily open them up from my messenger system in large packages.



I sketched over 15,000+ sketches in a matter of weeks and exposed myself to tons of different kinds of design implementations across different verticals. For example, the bottom navigation design was a common design pattern throughout many applications. The full study of this project will be linked here as soon as its available. Stay tuned!

One more thing, throughout this whole experience I went through 12 pens and way too much paper. Sorry, trees.