Anant Jhingran is no fan of the term 'data scientist.'
Currently the vice president of Products at enterprise API management specialist Apigee, Jhingran was formerly vice president and CTO of Information Management at IBM, focusing on the development of Watson. Watson, of course, is the cognitive computing system that in 2011 famously defeated former Jeopardy champions Brad Rutter and Ken Jennings, winning a $1 million grand prize.
"For the last few years I've been practicing what it really means to enable a large class of developers to build better and smarter apps," Jhingran says. "I've realized even more so that the data scientist has to go from being the nerd solving really hard problems to being the enabler of developers building apps."
"I've seen the transformation in myself," Jhingran adds. "I've gone from being the data geek focused on solving hard problems to seeing that my success is based on making other people successful."
Don't Call me a 'Data Scientist'
Jhingran says his discomfort with the term data scientist, which he first expressed in a blog post in 2011, just after giving a keynote at Hadoop Summit and a few months after Watson's big win on Jeopardy, is the result of a feeling that it sets people that do data science apart.
"It creates this aura that they're unapproachable," he says. "It also, in my mind, gives an easy way out for developers to say that data is very fickle and working with it is hard — 'Let's not bother with it. Let's not build apps that learn and understand and change.' To both sides the term is a disservice."
This feeling is part and parcel of a major shift that Jhingran says he believes is happening in the field of data science today as the capability to use big data becomes more mainstream in the enterprise and a key competitive advantage for those organizations able to make use of operational analytics and analytics for business intelligence. That shift is that data scientists are no longer magicians operating behind a curtain; they are beginning to work hand-in-hand with developers to deliver business value to end users.
"All the successful companies that leverage analytics see massive top line or bottom line improvement, but they see it because they've made these things mainstream," he says. "It's really got to be at that level of importance to make this thing succeed. Obviously technology is important and the data scientist has to evolve with it. If we agree with the fact that big data is going mainstream, in my mind there is one entity that sits between the work of the data scientist and the end user, and that is the developer."
Think Like a Developer
"The developers are the new kingmakers," he adds. "They are unlocking business value by building apps. The data scientist needs to have a new mindset — it's not just about solving big problems in isolation anymore. The mindset has to be: How do I enable these developers?"
For his part, Jhingran says he is working to drive that mindset at Apigee. Data scientists there are no longer in data science teams set apart from others. Instead, they've been spread out and now sit with developers in the lines of business.
"We made these data scientists actually sit in the teams that it is their job to enable," he says. "They live and breathe their problems. That has made a big difference in the data scientists' understanding that their job is to enable people to succeed."
The result is that these data scientists are now setting up their data products to be accessed by APIs that developers can leverage to power their apps.
"All that happens because the data scientists have not just done the hard work on difficult problems, but gone the next mile to enable the developers," he says.
However, he notes that developers, like data scientists, have to change their cultural mindset if they're to deliver the best value to end users.
"Developers have typically thought of themselves as programming either the UI or the app or the business logic," Jhingran says. "Whenever they talk about 'data,' they talk about data as persistent as opposed to data as analytics. It's not that they don't get it, it's just that it has always been difficult. We strongly believe that the developer of the future will not be a single-skill developer. Being able to play with data needs to become a very, very important developer skill."
The developer of the future, he says, will have to be multi-faceted, able to build an app in the morning, then build out an API in support of that app. Later that day, that same developer should be able to test the app to determine whether it is creating benefit, then put the insight gleaned back into the app.
"In five years, developers will have to be as comfortable playing with data as they are with business logic and UI logic," he adds.