KDnuggets Home » News » 2014 » May » Opinions, Interviews » Interview: Martin Hack, CEO, Skytree on Industrializing Machine Learning for Big Data ( 14:n13 )

Interview: Martin Hack, CEO, Skytree on Industrializing Machine Learning for Big Data

We discuss the mission of Skytree, product strategy, complimentary consulting programs, recent trends, and current expectations from Machine Learning.

Martin Hack Martin Hack has 20 years of experience, creating game changing technology products, services and strategies. His experience includes the management of product lines with revenues totaling $1.8B/year. He has launched ground-breaking products, driven world wide strategies and helped setting de-facto industry standards. As an expert on Trusted Computing, Virtualization and High Performance environments he became a sought-after advisor to many Fortune 500 companies and government organizations.

In his past roles he was involved in all aspects of the product life cycle including engineering, product management, marketing, business development and sales. He also developed and introduced products and services into virtually every commercial and government, channel and market segment for organizations such as Sun Microsystems, SonicWALL and GreenBorder (acquired by Google).

Here is my interview with him:

Anmol Rajpurohit: Q1. What inspired you to launch Skytree? Any particular reason for the name "Skytree"?

Martin Hack: At the time we (Alex Gray, my Co-Founder and CTO) realized that there was a) no general purpose Machine Learning System b) specifically one that does Big Data and c) pretty soon every enterprise will want to do what the Google's and Amazon's of the world have been doing for years.

Our name has two connections, one "Sky", our team has roots in large-scale astronomical data analysis and advanced data structures it just so happens that Skytree logowe have strong ties to the Astrophysics community (www.skytree.net/customers/seti-institute/), and "Tree" referring to the cosmic tree of knowledge found in many mythologies, a reference to our mission to bring state of the art research knowledge from leading edge science to the enterprise, such as machine learning. We were looking at probably 50-60 different names, and when we wrote Skytree down, we knew we had a winner.

AR: Q2. Why did you decide to build a general purpose Machine Learning system, and not one focused on a vertical, say finance or customer? What helps Skytree differentiate itself from the increasing competition?

Competitive differentiationMH: The latter would have certainly been easier. However, today we call Skytree a horizontal platform with vertical capabilities. What that means is that it is indeed a general-purpose platform but we have created vertical and application specific solution sets atop of the platform. Ultimately that¹s one of the reasons why customers end up working with us. The machine learning applications of today and the future are multi-faceted, a "one-algorithm-for-all" approach is not going to work for the Data Driven Enterprise.

AR: Q3. You launched Skytree Server in February 2012. Almost a year and a half later, you launched Second Opinion program in September 2013. What led to this shift from product-based company to a product-cum-service company offering advisory services? What was the motivation behind launching Second Opinion program? Skytree server

MH: At the core Skytree is a product company, and it¹s not so much of a change but rather a compliment to our offering. We realized organizations were looking at Skytree not just as a provider of software but their trusted Machine Learning partner. We wanted to help our existing and new customers that were just getting started as well as sophisticated users with a second opinion. To us it was logical consequence of responding to a market need.

AR: Q4. What is your favorite use case of Skytree product/services?

MH: Aside from our large enterprise customers, we are also working within biotech on early stage cancer detection. The impact of this type of work could be profound since it could end up saving many peoples lives.

AR: Q5. Machine Learning is more than half a century old. Why are we now seeing a sudden increase in its popularity, particularly beyond the academia? Where would you place Machine Learning on Gartner's Hype Cycle?

MH: Technology leaders have proven that it works and are having a ton of success with it. machine learningSo now everybody else wants do that as well.
In a way Machine Learning has become the new black gold. The application areas are literally endless and from where we stand we haven't even reached the inflection point. If you think about it, there are many industries out there that are just waking up to the reality of big data and data science.
I think the next 10 years will be dominated by machine learning (and its re-incarnations) and within 5 years every Fortune 500 will have a machine learning system at their disposal.

AR: Q6. Besides Predictive Analytics, what other applications of Machine Learning are of key interest to you? How would you define your primary target customers, those who will get the most value from Skytree's product/services?

MH: In addition to predictive analytics, the two that are coming up quite frequently are machine learning based recommendations and outlier/anomaly detection. This is where the power of a platform comes into play, customers usually want to do more than one thing i.e. to do classification followed by a regression or another machine learning task. Our target customers are usually large organizations with, well, big data. The more data the better ­ we can help with the machine learning part, but we definitely work with them in a collaborative fashion to establish the best possible use case.

AR: Q7. What changes do you see in the Advanced Analytics market since you launched Skytree more than two years ago? What factors would you consider as the key drivers of this change?

Drivers of changeMH: The obvious one is the arrival of big data, no big surprise here. In addition, customers are no longer satisfied with just using BI, which tells us what happened yesterday. They want to leverage things like machine learning to become much more predictive about their business. In essence they want to unleash all their data that they have been storing for years to help them get better insights and make the right decisions ­ faster. In other words they want to become a data driven enterprise.

The arrival of the data scientist and with it the Chief Data Officer/Chief Data Scientist is also happening. This may or may not be a centralized organization but it shows that customers are now putting much more emphasis on the results of their analytics and it's quickly becoming a massive differentiator from a competitive point of view.

Machine learning has the potential to change the game and market positioning for many organizations. It's one of the biggest disruptors over the last 10-15 years with true disruption potential, an opportunity for market leaders to become even stronger but also for new and traditionally overlooked organizations to establish themselves as a powerful player.

AR: Q8. Which book (or article) did you read recently and liked? What keeps you busy when you are away from work?

MH: Well, I'm somewhat of an information junkie and I'm a fan of KDNuggets. The article that most recently caught my attention was the White House Big Data Report. Outside of the office, most my time is spent reading but I am also a big music fan so every now and then I get to listen to some tunes and on occasion I will catch a soccer match. In another life, I was a professional soccer player for Bayern Munich in Germany and have been avidly following the news leading up to the World Cup Brazil 2014.