Kunal Jain — Published On September 11, 2017
Analytics Vidhya Beginner Interview Questions Interviews

“Corporate by chance, Entrepreneur by choice”

– Interview with Pankaj Kulshreshtha, CEO, Scienaptic Systems

If you have to learn experience of other people, I can recommend 2 sure shot ways of doing it:

  • First is by reading books. Books contain a ton on information about experiences people had in past and their learnings. Elon Musk, Bill Gates, Warren Buffet and many other leaders have been avid readers.
  • Second is by interacting with these thought leaders directly or indirectly.

Because of my role at Analytics Vidhya, I am lucky to have multiple opportunities of the second type. Today, I am sharing my interaction with one such leader – Pankaj Kulshreshtha, CEO, Scienaptic Systems

Pankaj has 20+ years of experience in analytics space. He holds a PhD in Quantitative Methods from IIM-Bangalore. He has grown multiple businesses in past and is still a researcher at heart.

Read on to know about his journey & story of setting up Scienaptic Systems.


P.S. If you have not bought tickets for DataHack Summit 2017 – this could be your opportunity to interact with thought leaders from across the globe.


Kunal: Hi Pankaj, thanks for taking time out for this interview. One of the fears I had when I was in my job was that if I stay too long in a job / secure environment, I might not become an entrepreneur later on.

When I look at you, you started your venture after spending close to 20 years in corporate. How did this journey start?

Pankaj: Every day we read or hear stories about a group of people who wanted to be Entrepreneurs right from the beginning. But, that was not my journey – In fact, even my entry into the corporate world was a chance encounter.

When I was finishing my doctoral work at IIM-B, GE advertised for a ‘Decision Science’ center and this was quite unique in 1997-98. At that time, predictive modeling roles and working with credit bureau data were unheard of in India. Even though my interest lay in academics, I decided to experience the corporate world for a while, and began with a catchy designation: Assistant Manager – Modelling. The experience of working in analytics was great, timing was just right. I did hands on work that still keeps me somewhat sharp in geeky discussions! The more interesting aspect of growth though was on other dimensions. GE is legendary for developing leadership skills and I was fortunate to grow in a very conducive environment. After about 4-5 years, it was pretty clear that I wasn’t going to academic world, at least not full time!

I am a very passionate believer that use of data and analytics will dramatically change not just the corporations but also our societies. Building and growing the analytics business at Genpact was very fulfilling and enjoyable, but I started getting a serious inner push to start thinking about starting a new venture in the last couple of years. I chewed on it and discussed the ideas widely. Came to a conclusion that there is much friction among technologies, processes and people that slows down the adoption of predictive analytics. Once that clarity descended, I finally decided to start on the entrepreneurial journey and here I am today – a very unlikely entrepreneur at an unlikely age!


Kunal: So does that mean your ideal clients are MNC’s who have huge amount of data and they may not be having the right infrastructure. Basically, they want to accelerate or transform the process.

Pankaj: Our target audience is big organizations, especially banks, financial services and insurance companies that have been investing large amount of resources, both in terms of monetary investments for buying new data infrastructure and also setting up machine learning teams.

But what we have observed is that in spite of making these investments for 5-10 years now, these organizations still struggle to deliver intelligent customer interactions. This is what we want to change, by enabling organizations to use and enrich omnichannel data and leveraging ML at industrial scale.

We have created a platform (“Ether”) that companies can use without having to spend tens of millions of dollars on long range projects but still deliver superior intelligence in their customer interactions. We work in such a way that we leverage the existing technology investments of our client along with our platform and accelerate the development of customer management strategies embedded with advanced ML algorithms. Everything we do is geared to produce significant business impact in short agile cycles of 2-3 months.


Kunal: In my experience, I have also seen companies making huge investments in data infrastructure and data science. But, I haven’t seen a marked difference from that in customer interactions. What do you think are the flaws in their strategies or common challenges your clients face and how do you solve them?

Pankaj: This is a matter of opinion to some extent. I have worked with global banks & financial companies in the past and I have seen a huge revolution in the way they use & manage data. I think there has been massive progress made over the years.

There is a plethora of young, Fintech startups that have dramatically improved the customer experience in the past few years. This kind of innovation in customer experience is hard to achieve for larger financial organizations because of a few reasons:

Larger organizations have multiple cross-functional transactions which makes it difficult for them to create a centralized & intelligent customer view at the speed of smaller startups. Hence, creating cross functional strategies to drive customer experience takes up a lot of time and effort in these organizations.

The other big problem organizations deal with is around creating a unified customer experience. e.g. If I reach out to a call center of a bank, they should know who I am, what my past transactions have been, whether I have visited the bank branch to resolve my issue etc. All that I (the customer) want is a unified customer experience irrespective of the channel. However, even as banks are getting better at building applications, the data and technology is not yet there to create a unified customer view

The third issue is the fact that most banks are still organize their P&Ls by products like mortgage, credit cards, loans, etc. Each of these products are separate business units, many times with their own data warehouses. But essentially, these separate business units are targeting the same customer / prospect pool. The challenge here, is to integrate the different business units across customer lifecycle to ensure appropriate product and price targeting so that customer feels an intelligent approach to how they are being offered and reached for new offers.



Kunal: How did you get your first set of clients?

Pankaj: Our very first client was an ex colleague who had gone on to become the CEO of a loans business. When he heard about our company, he wanted us to handle analytics and data science for his company. Then another client came in as a referral from one of my ex-colleagues.

To be honest, I have been fortunate. The people I have worked with in the industry had a certain level of trust in our ability to get things done and chose to partner with us. I believe that hard work never remains unnoticed; building relationships along your journey always helps.


Kunal: Can you tell us any use case which came in as a problem from client’s end and how did you solve it?

Pankaj: We worked with a large US cards company to help them manage their fraud. This client was experiencing surging fraud trends because of check-kiting and synthetic ID based credit bust-outs and the projection was incremental fraud losses of ~ $100MM. Historically they had used a set of rules to control the frauds, but these rules also had the flip side of killing a lot of good sales because of the high false positive rates. Within 6 weeks of our engagement, we designed a machine learning strategy for managing the losses and increasing the sales which were otherwise going on hold. We were able to deliver a clear impact of 20 million dollars in the fraud reduction and were able to reduce their False Positive Rates by 1/5th.


Kunal: How did you build your initial team?

Pankaj: I went about it the traditional way. I approached people from my network with whom I had worked in the past. I spoke to few of them and they decided to join us. Then the team approached their ex-colleagues & hired few more people.

Even as we hire more folks to build bigger teams, we still find that looking through our networks is the most effective way to hire the initial team.


Kunal: What was the most challenging time for you in the journey to build Scienaptic?

Pankaj: I started Scienaptic with a couple of friends and when one of my co-founders decided to leave the company after 9 months because it was not the right thing for them, the world came crashing down on me! I had to evaluate my willingness to even run the company. More than the loss of the co-founder, it was the loneliness in that situation that affected me. At that time, I did ask myself whether I really wanted to go ahead with my startup.

What kept me going was my belief in our idea. We were a team of 20 people at that time and I was totally convinced of the product as there was enough evidence that the market needed it. After that, some senior folks joined us, and since then, the company has been growing. Today, we have a great team and I feel energized by just being around them


Kunal: How do you define Scienaptic today, is it a product or services driven company and what is the culture like?

Pankaj: Before we hired any analytical folks, we hired engineers to build the platform. It was clear from day one that we were building a platform based company. The premise of starting Scienaptic was that the friction among technologies, processes and humans has to be dramatically reduced to enable ML to get embedded in organizations at scale and that needed a platform based approach. Our vision is still the same and we continue to make huge investments in developing our platform capabilities. Based on the serious traction that we are seeing, we feel our strategy is working out!

Our culture is hands-on and very tech savvy. Whole of our senior leadership is hands on and leads with personal example. We focus on building new things every day. We ensure that every piece of work that we do, delivers a significant impact to our clients. We strive to create a ‘future-ready’ workplace for ourselves, where there is an immense opportunity and flexibility to learn and grow, both professionally and personally. We are building a work culture which the millennials look forward to!


Kunal: Can you tell us more about “Ether”?

Pankaj: Ether is a platform that has data management, visualization, machine learning and workflow capabilities that come together to solve very specific business problems. We have designed Ether with several deeply technical innovations. It is a natural language based platform where working with structured & unstructured data is seamless. There is a significant push to democratize machine learning. Ether lets users work on complicated machine learning problems and evaluate the impact rather than worrying about writing thousands of lines of codes to solve the problem.

One of our core solution built on Ether is Credit Decisioning and Fraud Management. We understand that as technology evolves within Banks and Financial Institutions, there is an emerging need to also evolve the Credit Decisioning process and our solution is designed to do just that. Our advanced Machine Learning driven approach radically outperforms the traditional approach. This solution populates a rich multidimensional “Customer Consciousness” that can evolve to provide intelligent signals to other processes like customer service and collections.


Kunal: What are your top priorities today as the CEO of Scienaptic?

Pankaj: We intend to change the way organizations deploy and leverage machine learning in their customer engagements. As part of that vision, we are creating class-defining software that reduces friction among technologies, processes and humans. Another key element of our strategy is to make sure that everything we do delivers impact in very short agile cycles. Towards that we are continually innovating on our operating and engagement models.


Kunal: What are the different skillsets you look at while hiring in the US and in India?

Pankaj: We look for people who are hands-on and versatile with technology. Our engineering team has been based in Bangalore till now and our core developers are skilled in Java, Scala, Spark, Hadoop etc. As we deploy our platform to more US clients, we are building out an engineering team in New York as well.

The other team common to both New York and Bangalore comprises of people who have significant understanding of our platform and Machine Learning techniques. These folks primarily work with clients to understand their problems and create appropriate solutions using technology.

Kunal: This is kind of different than the regular practices in the industry. What was the thought process that went behind this?

Pankaj: The delivery & execution of work actually happens in both New York and Bangalore in line with what makes sense for the client. We don’t think of ourselves as an off-shoring company. We want to reduce the friction that can be caused in managing remote work. Our aim is to solve problems and transform the costs using technology and machine learning rather than off-shoring.


Kunal: Apart from India & US, are there any other markets where you have your presence or looking for expansion.

Pankaj: I am very conscious about geographical diversifications as we don’t want to thin out the intensity. Right now, we are focused on the US and UK as our primary markets. Most of our customers and prospects are Fortune 500 corporations and they have presence on both sides of the Atlantic. And because we have presence in Bangalore, we are also working with a few progressive Indian companies.



Kunal: You were one of the key members of the team which scaled up the analytics operations at Genpact. Tell us a bit more about that journey.

Pankaj: As I mentioned, I started with GE when analytics wasn’t known in India, about two decades ago. One good thing was that I started building models myself and that is the reason people still think I am technically solid. Over time the Analytics Center of Excellence at GE became popular across GE business units and I got to managing teams, started with 3-4 people team and by 2004 I was leading a team of about 250 people. At that time I decided to move into a functional role as Chief Risk Officer with GE Money, UK.  People would often ask me if it made sense to move from leading a team of 250 to leading one with 8 people! But, I think that was the most powerful career move I made because I actually got to see analytics live in action! What I learnt in that time & the relationships I built, helped me a lot in my journey. And living in another country was very enriching, it opens up our mind to the variety and possibilities.

In 2008, I came back to what had become Genpact. With hindsight, the timing was great as the meltdown of 2008 followed pretty quickly! Genpact was a tremendously rewarding and enjoyable experience. We grew 3 times while I was the business leader. I believe those six years taught me how to think big, scale organizations, and sell!


Kunal: These days machine learning, big data & AI is getting a lot of attention. How do you think things will change in the next 5 years?

Pankaj: I actually think all the hoopla around machine learning & AI is good. In my last few trips to New York, I have heard many conversations about Hadoop, Big data & AI on the streets in midtown. Unlike few years back big data analytics is being talked about a lot today. I think people still underestimate how much our lives will change as a result of big data and how insights will be generated and consumed.

I believe that enterprise engagement with customers in an interactive way through multiple channels must become a reality in the coming future. I also believe that Machine Learning will not just be a hyped-up curiosity. Rather, people will actually be using ML in real cases to solve real world problems.

The other thing that I think will change, is that more and more people will start getting comfortable with working with a set of technologies rather than just 1 or 2 primary enterprise platforms.


Kunal: What would be your advice to people already in data science industry & the beginners entering the industry?

Pankaj: There is a lot of talk around Big Data & Machine Learning in the industry. I believe that there is essentially only one problem that everyone is trying to solve – most organizations want to make sure their customers are happy and give them more business over time.

Today we have vast data available because of the digitization & one has the ability to do large computations on that data using machine learning & AI. There are numerous resources available to learn techniques and practice hands-on. My advice to people is to focus on solving business problems using machine learning – there is a scarcity of skilled professional who do that well enough.

I think there is no substitute for intensity & passion. Even if you have a niche skillset, you need to keep evolving your skills. We are living in times where ability to learn is lot more important than “knowing stuff”.


Kunal: Thanks Pankaj for taking time out for this interview. We hope to see you around in our community interactions in future.


Learn, engage, compete, and get hired!

About the Author

Kunal Jain
Kunal Jain

Kunal is a post graduate from IIT Bombay in Aerospace Engineering. He has spent more than 10 years in field of Data Science. His work experience ranges from mature markets like UK to a developing market like India. During this period he has lead teams of various sizes and has worked on various tools like SAS, SPSS, Qlikview, R, Python and Matlab.

Our Top Authors

Download Analytics Vidhya App for the Latest blog/Article

Leave a Reply Your email address will not be published. Required fields are marked *