I consider myself an accidental data scientist: Anubhav Srivastava of Viacom18 Media

Anubhav Srivastava is Head of Data Science at Viacom18 Media Pvt. ltd. Indian Institute of Technology BTech graduate, Kanpur started his data science journey with Evaluserve Consulting. He was also recognized as one of the top 40 data scientists under 40 by the National Machine Learning Developers Summit 2020.

In an exclusive interaction with Analytics India Magazine, he talked about his date with data science.

AIM: What drew you to data science?

Anubhav: My undergraduate thesis focused on the construction of a predictive model for the prevention of accidents in spacecraft. Around the same time, we lost Kalpana Chawla when Columbia broke down on their comeback. It opened my eyes and made me realize that data science can solve many real problems.

I started my career in equity research, where I worked on building predictive models to identify a company’s future prospects. It helped me understand how KPIs and metrics influence a business; why is data so essential to their survival, etc. This formed the basis of my interest in this field.

AIM: How important is it for aspiring data scientists to start early?

Anubhav: I consider myself an accidental data scientist. I worked in different fields before entering this field. Today, you find a lot of people transitioning into mid-career data science. People migrating from other industries have a strong advantage of having already been exposed to other facets of business. Additionally, the impact of data science on business outcomes is ingrained in these data scientists.

Having the right basics is very important. The advantage for newcomers is that there are many resources to help them improve their skills and motivate them to build models, make predictions, etc. However, a longer learning curve is better in the long run. The only suggestion I would give to aspirants is to not rush into this. Give yourself time and build solid fundamentals.

AIM: Is there a project that stands out in your career?

Anubhav: It would be hard to choose a project when your career spans so many industries. When you work in a new industry, you have to deal with two things. One is domain knowledge; the other is building a solution for problem solving. That said, there are two aspects of a project that give me an adrenaline rush. One is, projects where there are a lot of opportunities for human input. Trying to predict a model where there is a lot of human input is very difficult. The other is when there is a shortage of adequate training data.

I once worked on a project for a private equity client, and we built a model that predicted which global SME would be the next acquisition target. Challenges included: too many industries and therefore too many players and limited data accumulated over the past 20 years. Solving this kind of problem gave me a lot of motivation because we had little data to work with and a lot of human intervention.

AIM: How does Viacom18 harness data science?

Anubhav: We were a traditional broadcast media company; now we have a digital arm. It is a digital content company. Essentially, all data is stored digitally and all user interactions for anything built into the product happen online. So in that sense, business design is data-driven. Our business model responds to various aspects of day-to-day performance and the analysis of the data we derive from it. When it comes to data science and AI, we focus on two major aspects. Users and Our Content.

We have a two-pronged strategy. One is monetization. It’s an unforgiving world, and we’re trying to find ways to monetize our content while accommodating user demographic needs. We are constantly striving to improve the user experience to avoid making it invasive. This is where all AI/ML solutions come in. The main focus is recommendations. We’re leveraging AI/ML to monetize content that wasn’t previously monetized or to increase the likelihood of monetization: questions like how can you motivate users to purchase subscriptions, how can we incentivize people to watch more content, how can we get our users to watch more ads or click on said ads, etc. are what we mainly work on.

AIM: What are the main trends you see in data science?

Anubhav: First and foremost, data is now becoming an integral part of a business model. 5 to 10 years ago, the business model of a company was based on the consumer, the buyer and the intermediaries. No one would talk about data. Today, when presenting a business model to investors, stakeholders, etc., data and its monetization are discussed prominently. I think it’s a big change.

Early in my career, I had to convince people that data matters. Now everyone understands that. Another trend is the increase in data privacy. As far as I know, only banking institutions focus on data privacy and cybersecurity. Now they are gaining prominence in the organizational scheme of things. Another trend I see is no-code data science, which is essentially a cloud-native endeavor that allows plug-and-play models in an ecosystem. These companies are rapidly emerging in the country and around the world. I think it will gain tremendous ground in the future.

AIM: What is your long-term vision?

Anubhav: I’ve always had a keen interest in video. Whether it’s photogrammetry, adversarial networks or synthetic media. I think those are the areas that really pique my interest. I want India to be the home of innovations in video streaming and anything related to AI/ML in augmented or virtual reality, and I want to be a part of it.

AIM: What is your advice for data science enthusiasts?

Anubhav: Seekers should not let go of problems. Keep chasing a problem until you build something marketable or remarkable. This approach will help aspirants develop the mindset needed to succeed as data scientists. Experimentation is very important. It is very important not to think like a data scientist but like a problem solver. And being persistent is very, very important.

Sean N. Ayres