Fero Labs Led by EE PhD ’15 Alumnus Berk Birand Raises $9 Million To Bring Explainable Machine Learning to Industrial Manufacturing

By
ELIESE LISSNER
August 12, 2021

Fero Labs, the developer of the only Explainable Machine Learning software solely dedicated to the industrial sector, today announced the closing of a $9 million Series A round led by Innovation Endeavors, with participation from Deutsche Invest VC. This funding will support Fero Labs in expanding its product offerings to new sectors and ultimately push the industrial manufacturing community forward. The industrial sector has just begun to implement technologies into its processes to reduce waste and increase efficiency and profits.

Fero Labs was founded by Dr. Berk Birand (CEO), Columbia Electrical Engineering PhD ’15 alumnus and Prof. Alp Kucukelbir (Chief Scientist), formerly a Postdoctoral Research Scientist in the Department of Computer Science.

Dr. Berk Birand

Fero Labs focuses on bringing machine learning to the factory floor to optimize production, reduce waste, and improve product quality. As industrial IoT grows, automated assembly lines produce a continuous stream of data, but tools to analyze and act on the data are decades behind. Fero Labs’ statistical machine learning software provides factories with a virtual data scientist to predict quality errors and machine downtime. 

“I got exposed to the Internet of Things (IoT) space during my Ph.D. at Columbia. My colleagues in Prof. Zussman’s group were doing research on developing wireless protocols for low-power sensor nodes. This was long before IoT became a buzzword in the tech world. The potential impact it would have was clear. Given that I’ve always had an inclination for entrepreneurship, I decided to pursue a startup in that space,” Birand said.

“I initially started dabbling with the consumer side of IoT, but it soon became clear that the enterprise, and specifically the industrial angle was much more interesting, and the use cases were much more impactful. As we visited factories, we realized that most of the mathematical methods and tools they were using, like SixSigma and Lean, were based on classical statistics. More importantly, most of the modern advances had not made their way on the factory floor. This phenomenon, combined with the continuously increasing availability of data in the sector, created a very large opportunity,” he said.

Read press release.