Qualcomm Developer Network May Developer of the Month is Mark Jamtgaard from RetailNext located in San Jose, California. Mark is the Director of Technology, where he works extensively on computer vision and embedded hardware. He is currently working on the ongoing development of the Aurora v2 sensor, particularly around stereoscopic calibration and other aspects of the company’s computer vision solutions.
RetailNext, a Qualcomm Ventures portfolio company, was the first retail vertical IoT platform to bring e-commerce-style shopper analytics to brick-and-mortar stores, malls and brands. Through its centralized SaaS platform, RetailNext automatically collects and analyzes shopper behavior data, providing retailers with insight to improve the shopper experience in real time.
RetailNext introduced its first hardware device in 2016, the Aurora sensor, an all-in-one sensor integrating stereo video analytics, Wi-Fi, Bluetooth BLE and a beacon, designed specifically for the discerning requirements of the retail industry. New for 2018, RetailNext is introducing the Aurora v2, featuring not only the industry’s widest field of view, but also incorporating on-board, deep learning-based artificial intelligence at the edge.
Where do you get inspiration for your work?
I draw my greatest inspiration from my family, and in particular my two young children. It’s amazing watching them approach new situations with a truly inquisitive nature and very much a ‘beginners attitude,’ seeking to understand and learn. They’re such blank canvases and they’re painting with each and every one of their life experiences. That unquenchable thirst to learn and understand is something everybody in an R&D role should keep front of mind.
My technology heroes usually come from academia, and I gain a lot of value from the presentations and networking at events like the Conference on Computer Vision and Pattern Recognition (CVPR) and the Embedded Vision Summit.
What are some development tools and resources you can’t live without?
You can take a lot of things away from me, but please don’t touch my Emacs.
How did the Qualcomm® Snapdragon™ Mobile Platform assist in development of the Aurora sensor?
In developing the first Aurora, we quickly determined that the sensor, with its two camera lenses, resembled something that is already in the hands of most everyone – the mobile smartphone. Thus, RetailNext developed the first Aurora on the Snapdragon 805 Mobile Platform, maximizing the sensor’s video capture and output capabilities, as well as its processing efficiency.
The foundation of the Snapdragon processor delivered additional benefits when it came to developing the sensor’s software. By procuring existing development kits, software development started in parallel to the hardware design. As a result, several months were effectively cut off our timeline, allowing the entire process to be completed within RetailNext’s scheduling requirements.
Within the Qualcomm Technologies embedded business model, we had project-ready affiliates in manufacturing set up from the very beginning, allowing the contract manufacture of quality hardware in a minimal amount of time.
Based upon the resounding success of the original Aurora sensor, the development team decided to build the new Aurora v2 sensor by harnessing the power of the Snapdragon 820 embedded mobile platform.
Qualcomm and its portfolio of products, solutions and associates has allowed RetailNext to create truly ‘world firsts’ for its customers and their shoppers, and without, it would have been nearly impossible to replicate that feat within the same short timelines.
Where do you see the retail industry in the next 10 years?
Retail has fundamentally changed over the past few years. It started, of course, with the advent of e-Commerce, but it didn’t really take hold until the proliferation of mobile technology and Internet access put true power into the hands of shoppers. They have almost instant access to a wealth of information, and quite often they know more about a brand’s product and services than the brand’s own employees. Plus, they have an almost unlimited number of global alternatives to consider when making purchases. To succeed, Retailers need to be agile and respond nimbly to the needs, desires and values of shoppers.
Technology will continue to help retailers identify points of friction in their customers’ shopping journeys, and then proactively minimize them, whether its checkout lines, fitting room queues. inventory visibility and availability or anything else.
Artificial intelligence is making its way from online shopping into the store – the Aurora v2 sensor is one proof point, and it allows retailers to better understand how guests shop, and that knowledge goes a long way to providing more of what shoppers want, and less of what they don’t want.
As advanced as computer vision has become, we’re really just scratching the surface of artificial intelligence applications that train software to very accurately determine and categorize what it has just ‘seen.’ Applying those technologies to solve for practical problems and opportunities is why we come to work early and leave late.