Their new start-up, , lets shoppers click on an item they like, such as a shoe. Then, using patent-pending computer vision technology, the site pulls up a new page with similar shoes based on style and color.
Click on another shoe, and the algorithm does another iteration, further refining the search.
Modista searches through hundreds of thousands of items from sites such as Zappos.com, Endless.com, Nordstrom's and other online retailers. The website displays the results in a grid after each iteration.
"When you go to a mall you see a lot of stuff and the brain is very good processing images," co-founder AJ Shankar told The Industry Standard. "The intent of the site is to mimic that and enable them [shoppers] to discover products."
Shankar and his co-founder Arlo Faria are PhD students at U.C. Berkeley.
They hope to partner with retailers, and incorporate the computer vision functionality for individual websites. Zappos.com is their first client, and others retailers have contacted them, Faria told The Standard. At the moment, shoppers can also use the Modista.com portal to shop for watches, handbags and eyewear. Down the road, they hope to expand and include more product variety.
"We're looking at the big one, which is apparel," Faria said. "And other accessories like jewelry and home accessories."
During the 2008 U.C. Berkeley Business Plan Competition, which Faria and Shankar won, one of the judges joked that a team of male engineers trying to help a largely female clientele shop might want to invest in, well, a female perspective.
Yes and no, Shankar said.
"We both do rely on our girlfriends extensively in terms of the aesthetic design," he said. "Fortunately we're not tackling the much more difficult problem of understanding fashion and fashion sense. It's going to be a long time before computers are able to do that."
Faria and Shankar have taken a leave of absence from graduate school to work on Modista.