Building a Data Driven Organization with Srikanth Velamakanni

Nitika Sharma 28 Nov, 2023 • 11 min read

Analytics Vidhya’s ‘Leading With Data’ is a series of interviews where industry leaders share their experiences, career journeys, interesting projects, and more. In the 5th episode of the series, we are joined by a very special guest – Mr. Srikanth Valamakanni. He is the Group CEO, Co-founder, and Vice-Chairman of Fractal Analytics, one of the largest AI companies in India.

In this interview, he shares with us his insights and observations about building a data-driven organization. Being one of the pioneers in data analytics in the country, he also speaks about the changing landscape of AI over the years. Moreover, his deep passion for analytics, data science, and education is also highlighted in this talk with the Founder & CEO of Analytics Vidhya, Mr. Kunal Jain. Here are some excerpts from the interview.

You can watch the full interview here.

Join Analytics Vidhya Community to stay tuned with the upcoming Leading with Data sessions!

The Evolution of AI

Kunal J: I want to start with your very early days. You started Fractal 23 years ago when analytics was almost unheard of. So, you’ve seen this industry evolve from a very niche to where it is today. How has that journey been for you?

Srikanth V: The fascinating thing about our journey is that it has mirrored the journey of progress of AI. AI, as you all know, was a term coined in the Summer Conference at Dartmouth in 1956. I had access to the recordings and notes from the conference that they had in 2006, which was 50 years after the Dartmouth conference. Some of the attendees from 1956, like Marvin Minsky and a bunch of others, had attended the 2006 conference to discuss the progress in the world of AI, in those 50 years. And they were trying to understand what would happen in the next many years.

I saw the proceedings of that discussion and was fascinated because even in 2006, it turns out, people were actually discussing whether AI would go the route of first-order logic, by creating rules, exceptions, etc., or towards deep learning, or neural networks, as they were then called.

Earlier, when I was studying electrical engineering, the computer science department in my college used to learn AI. They had a course on AI, while we had one on neural networks – and they were two different things. AI meant rules – like fuzzy logic creating rules – and we were doing all these neural network things like fingerprint recognition, signature verification, etc., using very basic neural networks. This was in the nineties.

So even in 2006, the definitions and distinction was not very clear. And just 4 years later, suddenly, neural networks and deep learning popped up as core topics around the world. We started seeing impressive results from the labs of IBM, Microsoft, Google, and others. And then in 2011-12, something very interesting happened – Google realized that this technology was going to change the world in a very significant way. So they hired Jeffrey Hinton, who as we all know, transformed Google and added AI to each of their products.

So what I’ve seen over the years in terms of AI evolution is that there were the good old days. Then by 2010, the biggest companies that are AI natives or digital natives started realizing the potential of AI. And by 2015, the Fortune 100 and Fortune 500 companies from around the world, started to wake up to this. Around 2014-15, I started seeing many boards ask their CEOs to come and present their data strategy or AI strategy. However, it was still not a very big deal in India.

But in 2020, it became really big everywhere. Especially after COVID, it started taking off, and now in 2023, after ChatGPT, it has just become one of those things that we can’t stop talking about.

Fractal’s AI Journey

Srikanth V: If you look at the Fractal journey, the first 10 years of Fractal were all about problem-solving using analytics. So we knew a very clear decision problem. So we were seeing what is happening and predicting what is likely to happen using data, and helping companies make better decisions with it. Most of this was done using logistic regression or decision trees, random forest, XGBoost, etc. Until 2010, we were working with similar techniques on structured data for the most part.

By 2011-12, we created something called Fractal Sciences to explore the most impressive new problems in this field and invest in them. From this, emerged a bunch of products, and eventually, we hired Prashant Warrier, who helped us build Cure.ai and spawn it from 2015 to 2020.

Similarly, we spawned a whole host of AI startups like Crux Intelligence (AI-driven business intelligence), Eugenie.ai (AI for sustainability), Asper.ai (AI for revenue growth management), and Senseforth.ai (conversational AI for sales and customer service).

Between 2015 and 2020, we really took off in terms of using AI, ML, and DL to not only solve problems for clients but also create products and businesses. And since 2022, we’ve been focussing on building our own foundation models, our own diffusion models, and using them to solve all kinds of new problems that we never thought we could solve using data.

Turning Towards a Data-Driven Ecosystem

Kunal J: In between all of this, when was your Eureka moment; when you personally realized deep learning is essentially going to drive this data-driven AI, as opposed to the traditional rule and symbol-based AI?

Srikanth V: I’ve always believed in neural networks right from my engineering days. I always thought there was huge promise here.

But it was in 2012 or 2013 that I read a book called Super Intelligence by Nick Bostrom and then met with Lee Deng who was a very senior scientist at Microsoft. These two incidents made me realize what’s been going on in the industry for the last 3 years and how much I’ve missed out on. Meanwhile, even some of my friends were publishing papers on unsupervised learning and related topics. That’s when it hit me that I was already late to start, but I did start almost immediately. And that’s when we got Prashant, Suraj, and a bunch of others to join us and build a team. Now, in 2023, it looks like we started off about 10 years before most others.

Kunal J: How do you keep yourself up to speed with not only the business requirements but with the industry and research side as well? Do you have any advice for people who are leading these data science teams, in terms of how to balance the two?

Srikanth V: I think my advice to everyone in the industry is to get deeper – do your research and keep upskilling. Breadth is getting automated. We are moving towards a time where even if you’re good at client management, problem-solving, machine learning, people management, and multiple other things, and are very valuable today, you could become irrelevant tomorrow. It happened in the IT industry, it happened in the tech industry; it could happen in the AI industry as well. Your relevance now is a function of how much you have read or evolved in the last week, not just the last 10 years.

So my advice is if you don’t have the depth and the edge, it’d be hard to survive. So anyone from junior to middle to senior management, I would say do not lose the tech edge. Spend some time every day to read research papers and stay abreast of this tech because a lot is changing rapidly. For the junior to media level, I would say spend a ton of time reading the research papers and trying to replicate those results. For the senior people, If you don’t have that much time, at least work with somebody within the team trying to do the same. Also engage with the scientific community, researchers, and faculty and find out what’s the next big thing.

I still do that. Although it’s difficult to find time being the Chief of Fractal, I still manage to stay technically sound and updated to keep going and stay relevant.

Building a Data-Driven Organization

Kunal J: How do you see organizations evolving in the next 3-5 years? What would the organization of the future look like? How would the work be distributed between humans and machines?

AI is the most impressive productivity growth technology that we have ever come across.

– Srikanth Velamakanni, Group CEO, Co-founder, Fractal Analytics

Srikanth V: Let me start from the very obvious. Machines and machine intelligence will play a greater role in the way we work. It will bring in productivity growth and new intelligence and everything will become more automated. The machine component in the workforce will go up as it happened during mechanization, the Industrial Revolution, the information revolution, and now the AI age. Each of these phases has been about greater automation and the greater role of machines versus humans.

The second thing that’s happening is more intelligence getting embedded into it. Everything is just a tiny little bit smarter. This progress is compounding progress. So although in a short span of time, it looks linear, in the long run, say if you look at how smartphones have changed in the last 10 years, the growth is exponential!

So now, when it comes to work, I try to ask questions, like why should we work in the first place? This is a question that’s worth asking. What role does work play in our lives? And if it is about earning your daily bread, financial security, basic needs, shelter, etc., I hope that the world will get to a place where those are guaranteed for every human being. Then what happens is people would want to work for higher-order needs –  like love, passion, a sense of belonging, self-actualization, etc. Because we’ve already created so much wealth in the world, it no longer makes sense to have people work only to earn their daily bread. So in about 10 – 15 years from now, I believe, we will end up working because we want to and not because we have to.

Also by then, the organizational structure would change. I feel we’re still in the overhang of the industrial era where people are expected to work from 9 AM to 5 PM. Their in and out times are recorded and their every move is being watched. This will change from a time-based system to a knowledge-based system in the future. Companies would resemble an artist’s studio rather than a factory.

Fractal’s Problem Solving Approach

Kunal J: One of the key things that stands out about the problem-solving approach at Fractal is the combination of AI, engineering, and design. I’ve also noticed the focus you put on behavioral design and behavioral economics. So, when did you start moving towards that approach?

Srikanth V: This transition happened about 7-8 years ago at Fractal when I looked at where we were successful and where we were not successful.

We analyzed that and realized two things were missing. One was that we built something and it took many many months to implement that as we depended externally for the engineering power. While working with big companies like Tata and Airtel, although we would build the algorithms, it would take them a year or more to implement it into their systems, due to this, and we weren’t doing any of that engineering part.

The other thing we learned was never to take the client’s problem for granted. They might come to us with a downstream problem, let’s say ‘not enough credit cards are being sold’ or ‘customers aren’t using a specific feature as expected’ – but we must go upstream from there and trail it back to its root cause, which is the actual problem to solve. This is something that we learned very early on – to reframe the client’s problem and figure out what exactly to solve.

So yeah, that’s how we got to the recipe of great problem solving that requires AI, engineering, and design, and also the art of reframing the problem by working with the users. So around 6-7 years ago, we brought together a bunch of people from different disciplines ranging from sociology and anthropology to neuroscience and Java coding, and built a team for engineering, where they could all work together and create magic!

Srikanth’s Educational Initiatives

Kunal J: Another aspect that fascinates me is your passion for education. Even with such a busy schedule, you take time out to teach courses. You co-founded Plaksha University, and then you take a course every year going there. You also teach people internally at Fractal. What drives you to take that time out? And can you share some of these efforts?

Srikanth V: Firstly, Kunal, I’m very grateful to my teachers. Teachers have an enormous multiplier effect. You can create a lot of societal good if you improve teaching. If you take a very long-term view of the economy, education and entrepreneurship are the two vectors that will lead to better-educated people to solve problems, and entrepreneurs who are willing to take the risk to solve those problems. And together they can make the world a better place. Therefore, any time that I can spend in building that future society or future world, is hugely satisfying.

So founding Plaksha University was one of those things. But more than the money or anything else, I believe it’s the time and effort you put into it that matters. You must submit yourself to the cause. I may spend, you know, 40 hours teaching a course, but the 50 – 100 students who have now gained that way of problem-solving through this interaction, can go and change the world. Even one of them could change the world in a significant way. And I would take those odds because I think that through them, we can create a greater impact.

I teach a course called ‘Machine Learning to Make Better Decisions’. It’s a fascinating course that brings in different sciences, from neuroscience and behavioral sciences to problem-solving and cross-organizational processes for making. So the same topics, when I teach the students, are a little more technical and when I teach my Senior Manager, are a little less technical.

The hidden benefit of my teaching is that it keeps me young. It keeps me on my toes and helps me stay updated as well. Teaching is the best way of learning. You realize your own level of ignorance when you try teaching a topic. And then you invest time in learning the topic. You have a class coming up tomorrow then you better learn and you better burn the midnight oil in figuring out that topic. So it pushes me in the right direction as well.

Kunal J: To add to that, you recently launched the Fractal Data Science Professional Certificate on Coursera. Can you tell us a bit more about this data science certificate?

Srikanth V: I think it’s still in its early phases. The idea is to create a way to produce great data scientists in our country. The industry today has very few data scientists in India. As of today, there are 6 million professionals in the IT industry, and I feel like many of them need to learn AI and Data Science. Moreover, we’re graduating a million engineers every year, and they should know this. So the idea is to make a whole host of these people data science literate, with a much more fuller perspective of what it takes to use data to solve problems. The idea is to create a set of more fuller, more mature, more well-rounded data science professionals through this course, who are ready to go into the industry and solve these problems.

The Closing Round of Rapid Fire

Kunal J: As we’re heading towards the end of our session, we still have quite a few areas to touch upon. So let’s just do a round of rapid fire. What book did you last read?

Srikanth V: I try to read as many as a hundred books a year. So I’m always reading 20 to 25 books in parallel. But if I had to pick one book to recommend, it would be a book called ‘Influence: The Psychology of Persuasion,’ by Robert Cialdini. It is probably one of the best books I’ve read on any topic. Another book I always recommend is ‘How Will You Measure Your Life,’ by Clayton Christensen. And then there’s Viktor Frankl’s ‘Man’s Search for Meaning’. The book that I recently read, which I really like, is a book called ‘The Molecule of More,’ by Daniel Z. Lieberman.

Kunal J: If you were starting out today, what kind of startup would you build? What domain would it be?

Srikanth V: Great question. I wish I knew the answer. I will certainly not build a Fractal-like company. What I would probably do is take a very deep problem that is not solved, work backward to solve it, and then try to build a great company around that.

I think building one great company in life itself is a great fortune, and that too to the extent that we have spawned off a bunch of startups, which also have their own path to greatness like Cure.ai, Crux Intelligence, or Senseforth.ai. These are all very good ideas that could all become great companies in their own right. I think I already feel like I’m doing enough. So I’m certainly not thinking of a new idea on my own. But if at all I had to, it would be an idea I solve one problem, very, very well, and then build a global company.

Key Learnings From the Interview

  1. Artificial Intelligence (AI) was a term coined at the Summer Conference at Dartmouth in 1956. However, it was only by 2010 that neural networks and deep learning popped up as core topics around the world.
  2. Srikanth co-founded Fractal in 2000 to solve problems using data analytics. He later set up Fractal Sciences in 2012 which was more focused on using artificial intelligence for problem solving.
  3. Between 2015 and 2020, Fractal expanded to creating AI and ML-based products and businesses, giving rise to a host of AI startups like Crux Intelligence, Eugenie.ai, Asper.ai, and Senseforth.ai.
  4. Going forward, Srikanth believes that most functions in an organization will get automated, and only those who stay updated and have the edge can continue being assets to their companies.
  5. Srikanth is also passionate about education and has founded Plaksha University, where he teaches ‘Machine Learning to Make Better Decisions’. In this regard, he has also launched the Fractal Data Science Professional Certificate on Coursera.

Know More

Those were the highlights from our exclusive interview with the Group CEO and Co-founder of Fractal Analytics, Mr. Srikanth Velamakanni. You can watch the full interview here. Stay tuned to our ‘Leading With Data’ series on the Analytics Vidhya Community Platform for more exclusive interviews.

Nitika Sharma 28 Nov 2023

Frequently Asked Questions

Lorem ipsum dolor sit amet, consectetur adipiscing elit,

Responses From Readers

  • [tta_listen_btn class="listen"]