Today (Monday), Israeli startup Edgify, which automates Machine-Learning at edge, announced the completion of a $6.5 million Seed funding round. The investment was led by Mangrove Capital and Octopus Ventures. The round also saw a strategic investment from a yet to be named major chip manufacturer.
Accurately training ML and DL algorithms at edge
Edgify (formerly Pixoneye) develops technology that can train Machine-Learning between medical devices like MRIs, vehicle fleets, retail systems, mobile devices, and basically any device with an NPU, CPU, or GPU processing unit. The company was excited to tell that its ML training model is 99.98% accurate, mainly due to the fact that the models continue to perfect their learning from a constant flow of collected data.
The Israeli startup’s technology trains ML algorithms, which is usually done by servers with extreme processing skills, on edge devices like phones, tablets, and laptops. Essentially, the models utilize the endless data created from our mobile devices and use those same ML models to share and continue to teach the Machine-Learning deployed on our edge devices.
CEO and co-founder Ofri Ben-Porat explained in a conversation with Geektime what led the Ben-Porat and his partner to develop the technology in the first place. The story actually begins with computer vision technology developed under the name Pixoneye, which led to a meaningful discovery that urged the company to pivot.
Ben-Porat tells that while working on the previous company’s product, he and his team discovered that all the labor behind transferring large amounts of data to the cloud while complying with strict privacy regulations like GDPR are highly costly, therefore leading Ben-Porat to a distributed ML vision. “After a year of researching the subject we came up with a technological solution. Very quickly we learned that the distributed ML training models are a much better use of the technology we were developing, so we decided to focus on creating a platform for distributed ML as the product,” reveals Ben-Porat.
As with every conversation these days, Ben-Porat also explained how COVID had impacted the company’s abilities to complete the Seed round: “We were extremely lucky to have closed the funding round at the end of February just before COVID madness entered our lives. Afterward, we onboarded the strategic investor to our funding ranks.” According to him, because the tech is agnostic, it further freed the company to choose its target market, which was decided on retail in the end. The smart choice of picking supermarkets as the company’s main focus just prior to the food stores being the epic center of current day society panned out even though wasn’t planned.
Edgify was founded in 2014 as Pixoneye by CEO Ben-Porat and CTO Nadav Israel, and pivoted in 2018. The company notes that the funds will be put towards full global supermarket distribution of its ML models, as well as continuing research on the advantages of distributed Machine-Learning in the medical world.