5 Minutes with a CEO: Sejun Ra of ZEPL
What does it take to be a successful CEO, and where is analytics going from a CEO perspective?
- By James E. Powell
- January 6, 2017
Sejun Ra is the cofounder and CEO of ZEPL (previously NFLabs), which provides an end-to-end data analytics platform for all data types. Sejun founded NFLabs in 2011, which produced Apache Zeppelin, a Web-based notebook that enables interactive data analytics and provides one interface for data engineers, data scientists, as well as business analysts and decision makers.
Prior to creating NFLabs, Sejun spent ten years at Akamai Technologies, where he held several leadership positions in professional services, sales, and business development across multiple regions -- the U.S., Europe, and APAC. Sejun received a MS in Biology from New York University and a BS in Biochemistry from Hamilton College.
We asked Sejun about what most people don’t know about the job of CEO, which market trends are important (and which are hype), and where analytics is headed.
About the Job of CEO
Upside: What is the one thing you wish people knew about your job?
Sejun Ra: If I can point to one thing, it’s that there are a lot of things behind the scenes that I have to take care of that people are not aware of. Especially as a startup CEO, you have to make sure nothing falls through the cracks -- from finance and marketing to billing and contracts, even even ensuring the right Christmas party venue is setup. You are constantly going through a mental list to make sure everything is smooth. This is on top of hiring, fundraising, board meetings, ensuring great company culture, great product builds, etc.
That is not to say we don’t have people managing finance, production, and other roles, but at this early stage, the CEO is generally involved in all aspects of the business. I’m not sure if even my team knows fully everything that needs to be taken care of.
I remember a conversation I had with the chief of IT security of a bank. He said, “Nobody knows what I’m doing until something goes wrong.” I think it’s similar. Vision and products are still a large part of my daily job, but fundamentally I try to make sure my team is set up for success and that they only need to focus on building great products and nothing else.
What personality trait do you think CEOs need to succeed?
Humility -- it’s counterintuitive in this hyper competitive environment, but humility has taught me to ask for help instead of trying to do everything myself. It has pushed me to constantly keep learning; and especially because we try to hire the best and brightest, know that I’m not the smartest person in the room, and that’s okay.
Are you working on anything interesting right now?
Let me give you some background first before I answer your question. We’re starting to see some interesting trends in the enterprise space. On one end, the hyperconnectivity caused by the mobile revolution, as well as the Internet of Things (IoT), is creating an enormous amount of user data. This is making analytics invaluable -- a necessary team in every enterprise. It’s not just about hiring data scientists. Enterprises are building out full data science/data engineering teams with infrastructure and platforms to support them. On the other end, open source is becoming widely adopted by enterprises, often times being a requirement on solution requests.
These two trends are having a major impact not only on the technology used by enterprises, but on the profile of the people hired to manage these analytic workloads. These employees are very different today compared to days past. Data scientists today are not just mathematicians but engineers with working knowledge of deep learning frameworks. You no longer see job postings for DB admins. Instead, organizations want data engineers -- engineers with knowledge of various different compute frameworks.
So, to answer your question -- these people are in high demand, and as these profiles change, a new set of tools will arise to provide the velocity of pace they work at. It’s no longer acceptable for analytic workloads to take days or months. Modern-day analytics tools need to provide the ability to fulfill and visualize end-to-end analytics life cycles. This is what ZEPL is working on. Providing a single interface, a single analytic working environment, where full analytic workloads can be done AND multiple profiles of users can engage on it. We believe this will allow enterprises to realize their modern day analytics needs.
About Your Work
What trend or idea do you think is the most overhyped today in the corporate world -- be it about analytics or business in general?
I think the term big data is often misrepresented or misused. The size of available data has definitely increased exponentially, and will continue to do so, but good analytics is fundamentally about accurate and quick insights. The ability to quickly access, iterate, and analyze all data types is paramount, especially because analytics has moved from simple metrics to more complex behavioral questions.
What other trends are important?
Historically, there’s been a misconception that large enterprises do not do open source. Through the Apache Zeppelin project, we’re seeing the exact opposite. Large enterprises, across industries are adopting open source -- often times choosing an open source solution over equivalent closed source solution -- but we are also seeing a lot of contributions from these enterprises. From finance and health to aerospace and commerce, it’s clear that open source and Apache Zeppelin is filling a big need on enterprises large and small.
What is your favorite tool or technique that has made your job easier (and how is it easier)?
This is shameful self-promotion, but Apache Zeppelin is pretty cool. We use it extensively to track all our metrics, and as a dashboard to share internally. It’s convenient because we use it to analyze everything from small data sets on .csv files to large data sets by spinning temporary clusters on the cloud. We even use it to fulfill customer requests for further analysis and share it through our platform. Super easy and fast.
What’s the most common roadblock you hit in your work? How do you deal with it?
Hiring is probably our biggest challenge. We set a pretty high bar so even though we’ve been getting a lot of interest and resumes, only a handful make it through. It’s been frustrating at times, but we never want to compromise on people. I know it’s cliché, but we’re trying to make a deep impact on an established enterprise industry, so having the best and brightest is absolutely paramount.
Using Analytics In Your Work
Are there new technologies you want to try?
We’ve recently been looking more seriously at serverless architectures as well as some deep learning frameworks. We’ve already set up some prototypes and we’re pretty excited by what we’re seeing so far. Users won’t notice our use of serverless tech, but they’ll see greater support for deep learning frameworks on Zeppelin as well as implementations of it on our platform.
Where is data analytics/data science headed in the next few years?
Recently, we’re seeing greater investments on data analytics and data science. They’re becoming a much more important part of every business instead of “just a back office job.” However, because of the big data trend, much focus has been on quantity and size of data. I think we’re going to see greater focus on the following themes -- speed, access, and automation.
Speed is self-explanatory. More and more, the time between data collection and insight requirement is shortening. Enterprises are demanding real-time insights so decisions can be made faster.
By access, I don’t mean security, though security will always be a critical and important component of any data analytic platform. By access, I mean the availability for everyone to do data analysis/data science. Because of open source and the proliferation of tools such as Zeppelin, which makes data analytics/data science much easier and accessible, data analytics won’t be limited to just a select few “professionals”. It will become much more democratized, potentially open to even non-professionals.
For example, as a parent, what if you could have access to your children’s mobile apps data? By simply clicking on a data source (or multiple data sources), a parent can slice and dice the data to analyze their children’s usage patterns. Not by coding or writing complex algorithms, but by simply asking “Is my daughter talking with anyone I don’t know in Snapchat?” it would display all the names, times, and content. This is just one example, but you can see the potential applications of data analytics being available in our everyday lives.
In terms of automation, I’m referring to machine learning. There’s a lot of discussion around machine learning and deep learning use to gather deeper insights from data. This is important and will continue to enjoy considerable investment. However, a recent survey found that data scientists spend over 60% of their time on data cleaning/formatting, and it was also voted as the least enjoyable task. This is where machine learning can be utilized for great impact. Simply by optimizing this task, you already double your data science team, and this is where we believe Zeppelin can be a great interface to build such applications. It’s pluggable architecture enables quick experimentation, so application of different machine learning frameworks is easy and fast.
We’re living in amazing times where connectivity is pushing the boundaries of what we can know about things from the data we collect. The next five, ten years, we’ll see huge transformations in technologies and industries.