Computer Science Students Creating "Pandora for Fashion"

Its co-founder Tobias Lei calls it the “Pandora for Fashion,” and for those who just can’t figure out what to wear, it may be the ultimate time saver.

Lei, who is pursuing a master’s degree with CS @ ILLINOIS, along with fellow CS graduate student Gong Chen and creative lead Liz Li, a recent Illinois advertising graduate, will soon roll out an app called StylePuzzle, which helps make informed suggestions as to what to wear each day.

Liz Li, Gong Chen and Tobias Lei, co-founders of 'Style Puzzle' represent another start-up generated from Engineering at Illinois.
Liz Li, Gong Chen and Tobias Lei, co-founders of 'Style Puzzle' represent another start-up generated from Engineering at Illinois.

Here’s how it works. Login to Give it a few simple pieces of information like the weather and the occasion, and through its machine-learning algorithm, the app will provide photos of potential apparel combinations for you based on recent fashion trends and your specific preferences.

StylePuzzle is partnering with several outside brands and latest fashions are being added daily. The result so far is a photo library of more two million choices from more than 20,000 fashion brands. Much like Pandora, as a user makes a few choices, StylePuzzle understands more and more his/her preferences to more intelligently make future recommendations. However, what separates StylePuzzle from other similar apps is the ability to add items from your personal closet by either taking photos of them yourself or finding them on StylePuzzle’s extensive database.

It is also social, which means you can record what you wear each day, share it with your friends and in turn find out what they are wearing.

Lei is specializing in data science and has always had an interest in e-commerce even before internships with IBM and Amazon. The idea for StylePuzzle came after one of his friends routinely was having a problem with what to wear and would ask him to send the URL of some suggestions. After extensive research into fashion cloud sourcing, the idea became the subject of one Lei’s class projects.

Although he helps update StylePuzzle’s “inventory” daily to reflect recent fashion trends, Lei is quick to point out that he is by no means a fashion expert, merely a developer of an algorithm to help in the process.

“The suggestions are not based on my opinions,” he said. “They are based on thousands of people. There are other apps out there about closet management, but none we know of that offer a recommendation as well as have friends express their opinion by commenting and liking.”

According to its website, “StylePuzzle knows well about all fashion trends, and also well about the users themselves. This is why it can give relevant, practical, and high-quality outfit recommendations. We also want to differentiate with an approach to fashion that encourages people to wear the stuff they already own instead of chasing the latest products.”

StylePuzzle is currently beta testing it with a few dozen people and giving others a chance to sign up to have access to the app when it’s publicly rolled out sometime in May. In the first week of the launching StylePuzzle’s website, over 1,000 people signed up from as far away as the United Kingdom and Japan.

“Our early users are requesting some features to improve the shopping experience,” Chen said. “They want to add a wish list for things they want to buy as well as filter the navigation to make shopping easier.”

While the app has affiliate links to over 20,000 fashion brands on affiliate platform ShopSense, the long-term goal is to not only convince designers to advertise on the site, but be an on-line retailer for them. StylePuzzle would offer whole catalogs inside the app from only selected brands.

“We realize that in order to make this a big money operation, we have to be the marketplace where people can actually buy something from our site,” Lei said. “In order to do that, we need to further partner with manufacturers and brands and we’re heading in that direction. The advantage we have is we know their closets and thus capable of making smart recommendations,” Lei said.