I was actually at my first industry job with Area17 when I worked on projects for Facebook and Google. Area17 is an interactive agency in New York and Paris cofounded by Arnaud Mercier, one of the early prolific interactive designers. Inspired by pixel perfect design, we often brought this rigor to larger tech companies (like Google) to help them develop internal projects. I acquired a great foundation of knowledge around UX, strategy, development, and pixel perfect design while at Area17. At the same time I was cutting my teeth with big web projects like ESPN and Style.com.
After a few years there I moved to IDEO to work with a friend I met at Hyper Island (Stockholm interactive school) and to dive into a human-centered design. I had a desire to get much closer to the people we designed for which, as a UXer at heart, was very important to me. IDEO’s rich history always starts with human needs, and it’s a pleasure to be on the front line talking to people.
My work specifically was with Google Creative Labs, who would come up with ideas to pitch to various parts of the business. It was kind of an agency within the company. It was fun to work closely on projects that touched things, as Google consumers, we use all the time.
Just me! I had some help in getting feedback and testing from friends and coworkers but other than that it was a one man effort.
Originally I had very little understanding of machine learning, so the first “Khroma” was not very useful. I built a little tool to vote “yes” or “no” to a color (800 times), it took a really long time to train, and it would then produce 4 random colors that you like at a time. Once you refreshed the page it was gone. And that was it. I forgot about it for awhile.
It wasn’t until months later after lots of reading and experimenting with other machine learning projects that I realized how wrong I was doing it. I went back and spent many afternoons and weekends getting the algorithm right. Lots of time went to waiting for my computer to finish training neural networks. Once the AI component started working I began to build the UI around it.
I wanted the algorithm to create infinite pairs so I could pick the ones I liked. This was the inspiration for the UI design. I started making the pairs more interesting with typography and gradients. This part was the most fun for me, it seemed like the possibilities were endless!
The last part took awhile too. In order to launch it, I needed a landing page, a good onboarding experience, places for feedback and sign ups, etc. I also spent a long time making sure it worked on all browsers and didn’t have any bugs. I wanted the code to be clean and scalable. My friends started to think I would never launch it. “When is Khroma coming out?” they would ask. “Next week,” I would say. Then it wouldn’t. I didn’t realize how much work it takes to go from a tool that works for me to a product that works for everyone. So, I was very pleased when I could finally release it.
In the future, I would like to have user accounts so I can host the data in the cloud. This is far better and more secure than using local storage. I can also create the ability for users to create and share collections of colors. This will let me look at patterns in the data and make better algorithms too.
Nowadays there are so many information sources all over the internet. Where do you get new information about research, innovations and so on?
My top sources are:
Thank you! Despite being online all the time making Khroma and researching etc., I think we as a society are struggling with the way tech companies (social networks especially) make money. Primarily, they want to suck up as much of our time and attention as possible. This is a bad deal for us and is creating an unhealthy addiction to smartphones and the internet. My next project is an app to try to help with this issue by rewarding people to go offline. It’s very realistic so that’s all I can say now!
Интервью на русском доступно по ссылке.
У Дубаї відкрили магазин Nike з арабською в’яззю в дизайні
Behance представив редизайн профілю та сторінки проекту
Корпоративний месенджер Slack отримав новий логотип
В Канаді шрифт Calibri допоміг розкрити аферу
У Києві з’явилися точки збору шкідливого сміття та карта, на якій вони позначені
Shutterstock виклав 11 креативних трендів на 2019 рік