Subhasis Patnaik is the Senior Vice President and Global Head of Engineering at EdCast. Backed by top tier Venture Capital firms such as SoftBank and Menlo Ventures,EdCast is an award-winning, AI-powered Knowledge Cloud for Unified Discovery, Personalized Learning. The company has customers including Fortune 500 companies and multinational organizations.

EdCast uses artificial intelligence and its curation engine to bring together organizations’​ internal learning content, expert insights along with millions of external resources into an easy-to-use, personalized learning experience and knowledge platform.

I had the opportunity to do internship in Subhasis’s Data Analytics team last summer. Here’s an excerpt on an interview I recently conducted with Subhasis.

Can you tell us about how you use data science in your organization?

Subhasis: At EdCast, our core product is the “Spotify of Learning”. The way spotify uses a user’s preferences of genre of music and listening habits to recommend songs to a user, our recommendation engine uses the similar 

Large organizations typically have a lot of learning content – whether its content created within the four walls of the organization or from external sources (e.g., Coursera, PluralSight, Harvard Business Publishing) etc. It’s a daunting task (“information overload” problem) to navigate through the navigate through the plethora of content. For example, you are a software engineer who wants to move into product management. What kind of content do you need  to consume in what order in order to reach that goal? Can you emulate another person who has done that successfully before?

This is where we come in. Our recommendation engine uses key techniques like collaborative filtering. As such, data science is a key part of our product offering and is 

How do you distinguish between data engineering vs data science? Structurally do they belong in different organizations?

Subhasis: Sure. Given a problem space, data scientists are responsible for interpreting the data and figuring out the right solutions to the problem. For example, data scientists decide what models to choose from for our recommendation engine. The good news is that there are a lot of existing models to choose from. Do you start with content filtering, add collaborative filtering? Do you need to use deep learning models right at the outset? Data scientists dabble with these issues and decisions.

Data engineers, on the other hand, are responsible for building the infrastructure that helps machine learning engineers and data scientists to interpret  . For example, in our case, data engineers build the data pipelines that are needed to 

I think different organizations take a slightly different approach from an org structure perspective. For example, sometimes you see data scientists and data analysts being part of a different organization (e.g., Product) while data engineers and machine learning engineers being part of the broader engineering organization. Irrespective of where they sit in the organization, the collaboration among data scientists and data engineers is extremely important to build world class AI powered products.

What advice do you have for students currently in the data science programs?

Subhasis: Familiarize yourself. For example, there are a lot of good materials on how recommendation engines work.  

Participate in Kaggle competitions

Broaden your exposure. 

Look for internship opportunities.

Anna: Thank you so much for your time, Subhasis. I really learned a lot from you and your team!