Data science is coming out of the laboratory and into the boardroom. As innovative computing technologies such as cloud, serverless architecture, real-time streaming, and artificial intelligence mature, they are set to converge in a perfect storm of business opportunity.

“Where the impact on the business is happening is when you actually integrate AI in chatbots, in recommendation engines, in doing predictive analytics, in analyzing failures and saving those failures — and saving people’s lives,” said Yaron Haviv (pictured), founder and chief technology officer of Iguazio Ltd.

In a CUBE Conversation spanning the globe, Haviv spoke via livestream from Tel Aviv, Israel, with Jim Kobielus (@jameskobielus), host of theCUBE and lead Wikibon Inc. analyst, at theCUBE’s studio in Palo Alto, California. The conversation centered on the convergence of new technologies such as cloud, serverless, real-time streaming analytics, and data science and the possibilities that it opens for visionary enterprise (see the full interview with transcript here). (* Disclosure below.)

[Editor’s note: The following answers have been condensed for clarity.]

Welcome Yaron. You always have something interesting and new to share about what Iguazio is doing in the areas of cloud, serverless, and real-time streaming analytics — and now data science. Can you give us a broad perspective on the possibilities enabled by the convergence of those technologies? 

Haviv: Traditional analytics and even data science started in research labs. Now people are trying to make real ROI from AI and data science, so they have to plug it within business applications. So, it’s not just a data scientist sitting in a silo that generates some insights and rushes to the boss and says: “You know what, we could have made some money, a year ago, if we’d done something!”

That doesn’t make a lot of impact on the business. Where the impact on the business is happening is when you integrate AI in chatbots, in recommendation engines, in doing predictive analytics on analyzing failures and saving those failures and saving people’s life. Those kind of use cases require a tighter integration between the application, the data, and algorithms that come from the AI, and that’s where we started to think about our platform.

We worked on the real-time data, which is where when you’re going into more production environment and not data lakes, you need very good, very fast integration with data. We had this fast computation layer, which was microservices from day one, and that is allowing people to build those intelligent applications that are integrated into the business applications.

The biggest challenges I see today for organizations is moving from doing research on historical data and translating that into a business application or into impact on business applications. This is where people can spend a year. I’ve seen a tweet, saying: ‘We’ve built a machine learning model in a few weeks and now we waited 11 months for the productization of that artifact.’

Iguazio has been through several incarnations as a continuous data platform, intelligent edge platform, a serverless platform, and now I see that you’re a bit of a data science workbench. Can you connect those dots? What is Iguazio’s portfolio?

Haviv: They’re all nice marketing terms for this technology we’ve built. When we started, when we said continuous analytics, we meant feeding data in, running some of them, spitting some results out. This was opposed to the trend of Hadoop, which was a data lake where you throw data in, then you run the batch analytics, and a few days later you come up with some insights. So continuous analytics was a term that we came up to describe taking data in from different sources, crunching it through algorithms, and generating triggers and actions or response to user requests. That was unique and pioneering in this industry even before they called it streaming or real-time data science.

And now, if you look at our architecture, it is comprised of three components. The first part is a real-time multi-model database. The second is a microservice engine that allows us to inject applications of various kinds. We started with applications that do analytics; you know, grouping, joining, correlating. And then we started adding more functions into other things like inferencing image recognition, sentiment analysis, etc. Because we have this function engine, it allows us a lot of flexibility. Then the industry started calling this microservice engine serverless; we certainly were ahead on this serverless game.

The third element of our platform is having a fully managed platform as a service; a platform where all those microservices and data are managed through a self-service interface. Think of it as a minicloud. In the last two years, we’ve shifted to working with Kubernetes versus using our own proprietary microservices orchestration.

Having all the integrated stack created an opportunity for us to work with providers of edge. It’s not that we’re limited to edge, it’s just that what happens because we have an extremely high-density data platform, very power efficient, very well integrated, this has a great fit in the edge. But it’s also the same platform that we sell in the cloud as a service or we sell to on-premises customers so they can run the same things.

So, Iguazio is a complete cloud-native development and runtime platform. Serverless seems to be the core of your capability in your platform Nuclio, which is your technology. You’ve open-sourced it, it’s built for on-premises private clouds, but it is also extensible to be usable in broader hybrid cloud scenarios. Give us a sense for how Nuclio and serverless functions become valuable or useful for data science?

Haviv: We have a product called Nuclio Jupyter. A data scientist writes some code in a data science notebook and then clicks one command called Nuclio Deploy. Nuclio Jupyter automatically compiles his data science artifacts and notebooks, etc., and converts it into a real-time function that can listen not only on HTTP, but it can listen on streams, it can be scheduled on various timing, it could do batch, and so many other things.

If you think about data scientists, they’re not the best programmers, because they should be the scientists. So, by operationalizing their codes for serverless, you can cut back to market, you can address scalability to avoid rewriting of code, all those big challenges that organizations are facing.

Can you name some reference customers that are using Iguazio inside of high-performance data science workflows?

Haviv: We just announced a few weeks ago the investment of Samsung and Iguazio, which essentially has two pillars. One is that Samsung has adopted Nuclio as their serverless for internal clouds; and the second is that we’re working with them on a bunch of data science use cases.

Essentially, those are real business applications, at least three of which involve intercepting data from users and customers doing real-time analytics and responding really quickly. One of the things that we’ve announced is because of the use of Nuclio and some tricks that we’ve done within Nvidia, we have actually quadrupled Samsung’s performance.

If you don’t have AI incorporated in your business applications in a few years, you’re probably going to be dead. I don’t see any business sustaining competition without incorporating the ability to integrate real data with customer data and react based on that. 

Watch the complete video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s CUBE Conversations. (* Disclosure: Iguazio Ltd. sponsored this segment of theCUBE. Neither Iguazio nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE