Iguazio raises $24m to bring its Data Science Platform to new markets

Banking on the growing popularity of MLOps

by Max Smolaks 29 January 2020

Israeli startup Iguazio, which develops a data science platform capable of processing information in real-time, has raised another $24 million in venture capital.

The funding package comes from a combination of new and existing investors, led by INCapital Ventures, and brings its total funding amount to $72m.

Iguazio’s platform enables customers to build machine learning pipelines; it can aggregate data from multiple sources and uses Kubernetes for scaling.

“Customers are using our platform to build a pipeline that uses data efficiently, because at the end of the day, collecting data is a headache: you want to collect data from your real-time operational systems, collect historical data that is relevant to the users that are now transacting, put all of that into the model and do it in parallel will all the other users of the platform,” Asaf Somekh, co-founder and CEO of Iguazio, told AI Business. “What we enable people to do is to build that pipeline in an automatic way.”

Iguazio will spend the additional cash on accelerating the growth of the platform and expanding its reach to new markets.

The startup has also announced that it has signed up Payoneer, a payment company with some four million users, as its latest customer.

For the ops

Iguazio was established in 2014 to create a platform that would simplify deployment of machine learning at scale, while outperforming cloud-based tools from major vendors like AWS, Microsoft and Google.

Its data science platform fits into the emerging category of MLOps software, focused on the practical aspects of managing machine learning models.

MLOps tools are supposed to serve as a bridge between data scientists who develop the models in a lab environment, and operations teams that are tasked with supporting those models in production using corporate IT infrastructure.

“There’s a lot of hype, a lot of work going on around PoCs. And in many cases, the work in the lab where the data scientists have the data they need is very, very successful, they come up with great AI applications, with high accuracy models,” Somekh said.

“And then comes the hard part: taking that into production. This is where there is a very big disconnect between the perceptions of the people and the real world.

“People underestimate the complexity of taking the work that they’ve done in a lab and bringing it to life.”

Iguazio’s platform can be deployed on-premises, in the cloud, or on edge devices. Somekh said that roughly half of the company’s customers were in financial services, but its tools have also been used to create self-healing networks, and optimize ride-hailing services, logistics, and real-time recommendation engines.

The latest business to get on board with Iguazio is Payoneer, a developer of a global payment platform for business. It has used the automated pipeline to build fraud prevention capability that features predictive machine learning models, served in real-time.

“We’ve tackled one of our most elusive challenges with real-time predictive models, making fraud attacks almost impossible on Payoneer,” said Yaron Weiss, VP Corporate Security and Global IT Operations at Payoneer.