How to Create Machine Learning Products for Edge
Cloud has made hardware smart. It has revolutionized the decaying industry and made it sexy again. For some time. It did not take long for the users to realize that IoT products depend on the cloud so strongly that it actually limits their robustness. Such prosaic yet ubiquitous factors as intermittent network connectivity and limited bandwidth can make IoT devices rather useless.
These limitations have inspired a growing trend for Edge (on-device) applications that do not require constant cloud connectivity. Naturally, access to low latency real-time user-, device- and context data was destined to power a new generation of practical Machine Learning products.
But how do you build lovable and sustainable Machine Learning products for the Edge? In the talk, I will delve into the best practices of creating ML products for both consumer and industrial Edge platforms.