When Ashwini Asokan and Anand Chandrasekaran moved back to India about a year ago, the couple had a singular obsession—how to make machines more useful by making them a bit more human.
“The idea was to make devices smarter by giving them the ability to see,” said Asokan, a Carnegie Mellon graduate who had previously worked on user experience and interaction at Intel.
So, along with Chandrasekaran, an IIT-Madras engineer with post doctoral experience at Stanford University, Asokan setup their own firm, Mad Street Den, with an initial capital of $50,000.
At the heart of the startup is a cloud-based platform, MADstack, which can be incorporated into applications under development or even existing products, which need a dose of artificial intelligence.
That means a mobile phone camera can be harnessed to detect different parts of your face, picking up on gestures like moving your head from side to side and even expressions and emotions.
But what is most useful—and perhaps, most lucrative—for India’s fast-growing e-commerce market is its ability to recognize objects.
“If someone takes a picture of a top, our platform can help identify different things about the top, like if its pink, what kind of cut and design it has and then our technology will pick up the closest match on an e-commerce site,” Asokan explained.
For an e-commerce industry that’s projected to grow from $10 billion to over $40 billion in five years, and already has 11 categories and 42 players, Mad Street Den’s offering could be disruptive in a market that needs differentiation. Any app developer could make an app with Mad Street Den’s technology powering its back-end.
Brick and mortar retailers, already frustrated by customers walking into their shops to look and feel products only to later order them online, will resent this platform if it were to come into widespread use.
An undisclosed Indian e-commerce platform is their first customer, which is likely to launch their object recognition feature in October. Besides e-commerce, the startup is also looking to focus on gaming, social media and analytics.
But none of this is exactly new. ‘Mobile image recognition’ has been identified as a potential game changer for sometime now, and global e-commerce players like Amazon and Ebay already offer similar applications.
Mad Street Den’s pitch for their own technology is straightforward. Instead of multiple pieces of infrastructure for different functions, their MADstack platform offers one common solution. It helps that it also requires low computing power and has scalability across devices.
And more importantly, the founders claim, it could have significant commercial potential simply because the most basic component of the platform is something that users are already doing—and willingly.
“People are used to taking pictures of things they like using their camera,” Asokan said. “Unlike wearable technology, this doesn’t need a change in behaviour for users.”