Om Deshmukh: Mirroring value via Data Science

Data Science is the study of origins of data, what it represents and how information/value can be generated from it. Dr Om Deshmukh, Senior Director Analytics Solutions & Data Sciences Innovations, Envestnet | Yodlee, is an expert in theoretical and applied Machine Learning & Data Science and has over 50 patents filed/ granted and over 50 international publications in his name. Om is recognized as one of the top ten data scientists in India. He initiated and led IBM’s billion dollar global strategy on “Analytics for Personalized Learning”. He spoke at length with Corporate Citizen on what Data Science is all about, how Data Science can also play a key role in understanding human behavior, how to generate value from data and how to avoid misuse of data

It goes without saying that the togetherness of two decades comes from shared interests and common priorities-right from their student days at the University of Boston (where they first met.) All of 40, this dynamic couple is, nevertheless, a set of two very individualistic spouses: As senior director, Analytics Solutions & Data Sciences Innovations, Envestnet | Yodlee, he is one of the country’s best-known scientists, while she is pursuing a PhD in Biotechnology from Bangalore University. She also has a Visiting Student Fellowship in Nanobiophysics at Raman Research Institute.

Even as Om Deshmukh and Lavanya Muthukumar double up as proud parents to Devdatt and Vasundhara, they are pretty clear that both professional triumphs and personal happiness are rooted in a singular pursuit of what makes you happiest as an individual. So whether it is Om learning drums or Lavanya quietly running “Solace”, her support school for children from underserved backgrounds, this is one couple that has their own way to balance the yin and the yang

Tell us about your education and career.

Om Deshmukh: I did my B.Tech in Electrical and Electronics Engineering (EEE) degree from BITS Pilani. After my graduation, I went to pursue my Master’s in Speech Recognition from Boston University, USA. Two years in, my Professor, Dr. Carol Espy-Wilson, who had moved to the University of Maryland College Park, Maryland, USA, gave some of the students an opportunity to join her at the University of Maryland College Park. I was one of the two students who received the opportunity. There, I started my PhD in Speech Recognition, Speech Enhancement and related topics. The underlying theme was Machine Learning and how data-driven insights can be generated, but at that time, I focused on building Machine Learning algorithms specifically for speech and audio data.

After which, I did a post-doc for about nine months from the same university but it was a joint project between the Electrical and Computer Engineering (ECE) department and the Speech and Hearing Sciences Department (SHS). The idea was to use speech enhancement techniques that we had built as part of my doctoral work for people with hearing impairment.

After completing my post-doc I moved back to India. IBM had made me an offer to join their India research lab in New Delhi. I joined there as a Research Scientist and worked with them for six years. One of my substantial contributions at IBM was to drive a Global Technology Outlook (GTO) topic in 2012. GTO is an annual outlook of the technology-driven future which has a global impact for IBM, its partners and its clients. I had to develop a pre-strategy and it was a competitive situation because everybody believed that their piece of technological outlook was the next big strategy. In the end, only five got selected and mine was one of them. The topic of the presentation was Analytics for Personalized Learning. I presented that topic to IBM CEO and all her directs; I would call it as the biggest learning experience of my professional life so far. It was a 45-minute engaging discussion wherein, I was able to get a sneak peek into how a Fortune-500 CEO thinks at the strategy level. I had a lot of help from executives and global team members. It was not just a professionally fulfilling experience but also a tremendous learning experience. I learnt how technological transformations get combined with the changing business environment to formulate strategic strides. Over the next few months, we initiated several projects to make sure that the strategy got implemented.

In 2013, I joined Xerox Research where I drove the Multimedia Analytics Research Strategy. My responsibility was to drive technology and research in multimedia analytics. We were doing that as a horizontal capability, which means without any particular business vertical in mind. Later, the business vertical to which we applied this technology was predominantly the education sector. That was also the time, 2013, when MOOCs (Massively Online Open Courses) were coming up, for example, Coursera, Udacity, among others.

Close to home, NPTEL, (National Programme on Technology Enhanced Learning) had a lot of educational content recorded from the classrooms of IITs and IIScs, which were made available on the NPTEL channel on You- Tube. We applied our multimedia analytics algorithms to these videos and summarized them so that students who were searching for a particular topic were shown the content that was most relevant for them. For example, we considered simple things like accent mismatch between the student and the lecturer, to more complex things like the required technical background, to even more complex things like the likely end goal that the student has and the content sequence that the video was part of.

In 2016, I very strongly considered venturing out on my own in the ed-tech space. But I got cold feet at the very last minute. Then I thought a monthly salary was more comforting at that time, as I had financial obligations. My ex-boss (who continues to be a great mentor) and some of the then team members built the startup. Around that time, I left Xerox and joined Envestnet|Yodlee. My intention was to use the Data Sciences research – which I had built over the years at University of Maryland, IBM and Xerox and apply it in actual product delivery setup, where an end-user uses it. I decided to get into the Fintech industry and that’s when Yodlee happened.

I joined Yodlee as the Data Science Head to run the delivery of their Data Science products. Yodlee provides solutions to financial institutions, those institutions in-turn help their clients with personal financial management. Many a time, it happens that people don’t have the time to look at all of their financial transactions and make sense as to what are the ways from which the money is coming in and what are the ways in which the money is going out – inflow and outflow of finance. Yodlee provides these solutions to financial institutions. For example, at Yodlee, we will help customers identify how regularly they are spending money on various spending categories such as restaurants or entertainment, i.e., provide the user with an easy view of their financial behaviour.

Why did you transition from Engineering to Data Science?

In fact, there is no transition as such. Engineering is a core component of Data Science. All of what we learnt during the engineering years from Linear Algebra, Probability and Statistics, Operations and Research, Optimization, all of that is the crux of Data Science. When you apply it to a particular vertical, that is when it comes into the popular media but the underlying technology is still computer science and engineering. For example, for my PhD work, we had used a technology called Hidden Markov Models (HMM); the underlying thing in HMM is dynamic programming and dynamic programming comes under Computer Science. These are the foundations of Data Science.

"One major way to stop data misuse is to spread awareness about the value of data. Data has to be shared for value to be generated, so we don’t want people to go to the other extreme and not share any data"

Tell us about your role as Senior Director, Analytics Solutions & Data Sciences Innovations at Yodlee.

My role is to make sense of the different kinds of financial transactions that flow through our system. I gave the example of showing the financial mirror to the customer, my team drives a lot of that work-taking raw data and generating insights which are beneficial to the customer.

Do you think data gets misused and what should be done to stop it?

One major way to stop data misuse is to spread awareness about the value of data. Data has to be shared for value to be generated. We don’t want people to go to the other extreme and not share any data-it will not generate any value. At the same time, people should be aware as to where the data is going, how the value is being generated, and where the ownership of data lies. Companies should have very rigorous checks and balances every time data moves from one setup to a different setup. No matter where the data is moving to, there has to be a good set of checks and balances when data moves.

In my opinion, data protection has to be a continuous process. At no point can we declare victory stating that our data is completely safe. Because what happens today may be safe, but tomorrow, opposing elements may try to break what you have created. In that sense, my strong belief is that the process of ensuring that the data is safe has to be a continuous process, which is crucial.

How will AI / Machine Learning / Data Science help financial services?

Financial wellness is critical. Data Science should be applied, in the field, to generate value. Today, people earn and spend money without necessarily pausing to appreciate the significance of the financial transaction – this is going to be more and more frictionless with auto-debits, auto regular debits and more. It is good because it saves time, which is a welcome move. But at the same time, financial wellness becomes much more difficult when you are not consciously approving a particular bill-there is no control. That is where Data Science becomes very important-to create a holistic solution to provide a view of the inflow and outflow of your money. And then provide some insights on top of that. For example, a customer may be spending more money on food or movies or travel and so on. In the end, let the customer decide whether it was intended or not. Data Science would bring in great value in the space.

Will AI replace Data Scientists?

Hmm…interesting! I think there is scope for AI to help humans be more efficient. Data Scientists will be critical to realize the full potential of AI.

What you think is the future of Data Science?

Even with all the rapid developments happening in the area of Data Science, I believe we have only realised about 1% of the true potential of Data Science. In the near future, I hope Data Science will fully automate several of the repetitive tasks, alert human experts when anomalies are detected along with a likely root cause analysis of the anomaly.

"My biggest challenge is spreading awareness about Data Science and making sure that expectations are realistic. Data Science is not a magic wand which will solve all of your problems. It will solve many problems and in more systematic ways that no other techniques can do"

Are women adequately represented in the Data Science field?

There are a lot of positive movements happening to make sure that women are adequately represented in Data Science. It is happening across all different sectors, not just corporate alone, but government, educational institutes, small industries, who are doing quite a bit to make sure that awareness is spread, and representation is adequate. On a bit of a personal note, I dedicated my PhD to the three women who have influenced my personal and professional life profoundly: my mother, my PhD advisor and my wife.

How well-known is Data Science as a field in India?

Data Science as a buzzword is known, but as a field, it is still not understood well. Just like the phrase that says half knowledge is dangerous, similar is the status of Data Science in India. People think of Data Science as a Panacea. Data Science can solve a lot of problems but with certain caveats. Data Science has to be applied in a systematic way, the rest of the eco-system has to line up and so on. While Data Science has power and potential, what is happening is that it is being used as a buzzword to demand the impossible. As I had said earlier, spreading awareness regarding the capabilities of Data Science is essential.

What does it take to be a Data Scientist?

The term ‘Data Science’ is not very well defined so there is no clear answer to it. I will state four aspects of Data Science.

Building state-of-the-art algorithms-that is one aspect of Data Science. For that, one has to be very good at Maths: Linear Algebra, Probability and Statistics, at least these two streams and adjoining streams like Calculus and so on. This applies if you want to build state-of-the-art algorithms.

The second aspect of Data Science is when you want to use algorithms that are already established but now you want to scale them, you want to make sure that they run in real-time or they run for a particular real-life problem that you have. That needs slightly different skills, obviously, understanding of the algorithms but also the skills of scaling a particular algorithm to big data volumes. Which means you need to have good engineering skills, you need to be aware of the latest technologies, the big-data technologies are evolving quite rapidly. No one can say that I understand every new technology. But one must know the technological components that they need to build, whether to use Hadoop or Scala or any other. These frameworks should be known. Somebody who goes into the depth of algorithms may or may not be able to go in the depth of these scalable technologies and that is alright, you don’t have to.

The third aspect is what problems are worthy of solving? This is a big thing to look at. I may solve a problem using the latest technology and make it highly scalable, but, if it has no value, it is worthless. One needs to have the business acumen to figure out whether the problem is worth solving or not.

The fourth aspect is making sure that an adequate amount of relevant data is being captured or not. And identifying these data gaps.

What kinds of challenges do you face in your career and how do you resolve them?

My biggest challenge is spreading awareness about Data Science and making sure that expectations are realistic. Data Science is not a magic wand which will solve all of your problems. It will solve many problems and in more systematic ways that no other techniques can do-I stand by that and I believe in that. But there have to be realistic expectations, Data Science has to have an eco-system around it to perform.

Do you face any challenges when you deal with millennials?

It would be wrong to say that there are problems specific to a particular set of people. We cannot paint anybody with such broad brushes. Society is evolving very dynamically and hence the kind of problems that you will face will also be a lot more dynamic.

What is the philosophy of life that you live by?

Take a long-term view of everything. Secondly, there are no silver bullets in life.

What is your idea of relaxation?

To spend time with my family. Interacting with different sets of people, because there is always something to learn.

What are your hobbies?

I like to read. I read a lot of biographies; books about the companies that have made it big in the industry.

Lavanya Muthukumar: ‘Curiosity, the cornerstone of all commitments’

Walt Disney famously said, “We keep moving forward, opening new doors and doing new things, because we’re curious and curiosity keeps leading us down new paths.” Lavanya Muthukumar would agree. Even as she hopes to complete her doctorate “any day now” in the cutting-edge field of Nanobiophysics, she points out that while love and passion for one’s subject are a given, the world of research is both powered by and rooted in that single word-curiosity. “It is what brings you back to the table day after day,” she says simply.

Coming from a family that is heavily into academics, the pure sciences were not an unusual choice of subject for her. With Biology as her core subject, she went to acquire an M.Phil. followed by a doctorate, whilst holding various research positions along the way. As of now, she’s in full-on student mode, though she admits, “It’s the hardest thing to do because it entails putting your ego far, far away.”

But even as she balances her studies with the upbringing of her two children, she is also passionate about giving back. Precisely the motivation that has her running a small but efficient support school for children from a spectrum of underserved backgrounds. Unassuming, low-key and affable, here’s more about Lavanya Muthukumar, in her own words

What made you opt for a subject from the pure sciences in the first place?

Lavanya Muthukumar: Coming from an academically inclined ‘Tam Brahm’ family in Bengaluru, it was a given that I would study beyond bachelors. Education had been a top priority for even my mother’s generation and pursuing a post-graduate degree in either science or English was a popular option.

From the beginning, I would score well in Biology, so I presumed it was my subject. My mother, being a doctor, would have preferred me to be one as well. But each individual has their own temperament. Not very driven, I was happy to cruise along, unlike possibly, the students of today who are far more focused than my generation was. (smiles)

So I a pursued a bachelor’s degree in the subject followed by a Masters from St. Joseph’s, Bengaluru-following which, I did a stint at a lab in St. Joseph’s. Around this time, I took my GRE and decided to go abroad to the Boston University where initially, I intended doing my PhD, but due to various circumstances, did an M.Phil., instead.

That’s where Om and I met. Post marriage, I worked for three years at the University of Maryland in a research position. Post the birth of our son in 2006, I took a break, following which, we were in Delhi. Here, I worked at the National Institute of Immunology. On our return to my home-town Bengaluru, I thought it would be a good idea to complete my PhD, as I finally had support from my mother and mother-in-law.

How easy is it studying at this age?

To be honest, it’s the hardest thing I have done, learning how to learn once again. Being a student requires you to put your ego somewhere very far, far away, but on the other hand, it’s hugely rewarding. The pursuit of knowledge is lovely and so is the joy of learning from people half your age. As a doctorate student, one is expected to teach masters’ or graduate students at a university, but you end up learning so much from them.

Apart from being enlisted for your PhD programme, you are also affiliated to the Raman Research Institute. Research and academics are far from a glamorous pursuit. There is a lot of back-breaking labour involved, to put it mildly. So what keeps you going day after day?

Love of the subject is important, yes, but it’s a given. To keep going back to the drawing board, day after day, you need to be driven by curiosity. Does the project on hand interest me? What is it teaching me about things I don’t know? Is it pushing the envelope in an unexpected direction? The curiosity as to where the answer might lie is what keeps any researcher going.

You are pursuing a PhD in Nanobiophysics-quite a mouthful, apart from being a very cutting edge field. Please tell us more about it.

Well, Nanobiophysics is a new branch of science that operates at the interface of physics, biology, chemistry, material science, nanotechnology, and medicine. And it is an exciting new frontier. Simply put, you study biological molecules or cells at a level that wasn’t possible before.

Thanks to the special instruments available today, you can study every interaction within a cell, how two cells bond and interact with each other and so on. As you know, each and every living being can be defined in terms of the cells and thus these nanoscale measurements for proteins, molecules and biomolecules are valuable both from a pure science and medical point of view. The sheer depth of the study that wasn’t possible before is significant from the former viewpoint, whereas the latter is important for the medical advancements we hope to make. For instance, the study of cancer cells and how they move. All of this is telling us things about the behavior of molecules that we had no idea of before.

You have been running ‘Solace’, a support school for children from underserved backgrounds for over a year now. To that end, you employ teachers as well. Can you tell us more about it?

Solace is an idea that has its origin in an episode that took place a couple of years ago. A lady living in the same building as us asked me to help out her daughter with English-a language that put her at a significant disadvantage. Keen on helping her out, I learnt a special phonetic method to help her learn. The outcome was amazing, very quickly she transited from scoring 3/10 to marks in the realm of 8 or 9/10. It also renewed her interest in learning just in the nick of the time, before she lost confidence in herself and hope in the system.

In short, “Solace”, the support school I run, is meant for kids from a spectrum of underserved/tough situations. Constrained finances are only half the story; there are plenty of bright kids who come from backgrounds that do not have easy access to English. Considering that English is a core medium of instruction, not knowing the language can put even a bright child at a significant disadvantage. As of now, we have eight kids in the primary section and three in the secondary section. Typically, they come to Solace before going to their regular school.

“Solace”, the support school I run, is meant for kids from a spectrum of underserved/tough situations. Constrained finances are only half the story; there are plenty of bright kids who come from backgrounds that do not have easy access to English"

Family above all. Om and Lavanya with
daughter Vasundhara

Tell us how you met Om.

Well, we met at Boston University. But while he was studying Engineering, I was pursuing Bio. We got talking through the Indian Students’ Association- Tarang. It was a relatively small community that we were part of and so getting to know people was easy. Tarang, for its part, fulfilled a really important role, in helping Indian students acquaint themselves with local surroundings. Simple things like opening a bank account or knowing where to shop can make a huge difference in those confusing first days and here’s where Tarang lent a helping hand. In the second year of my involvement with Tarang, I was elected as the President, and Om became the Treasurer. That’s how we really started interacting. Taking decisions on what to do next and how to spend our funds were a responsibility but we enjoyed handling it. So a friendship gradually came to life and as time went on, we thought perhaps marriage is a logical step, considering we got on well. Of course, our parents had some misgivings. He’s a Maharashtrian Brahmin from Aurangabad and I am a Tamilian Brahmin. But we overcame those worries and they gave us their blessing.

After so many years, what would you say is your understanding of the institution of marriage?

Typically, marriage turns out to be a completely different ball-game from your life as a youngster. (smiles.) The gap between expectation and reality is often quite a huge one but one has to be willing to change, accept and compromise. For example, I lived independently for quite a bit-sharing an apartment with other girls pursuing their PhDs in Boston, where a neat time table and efficient system of co-operation took care of our lives together. If only I had known I am going to be responsible for cooking for four, for the rest of my days (laughs), my preparation would have been different. Typically, after kids are born, the mother, by default, becomes the primary caregiver and that is that. But marriage is also rewarding and it definitely has its moments. The correct attitude helps. Over and above everything though, for a marriage to really fulfil its purpose, both individuals must first see to their own happiness and prioritise it over all else. While getting on together is a good thing, make sure not to lose yourself in the process. It may seem like a paradox, but I do believe it is hugely important to love yourself before you love your spouse. Your inner happiness must be in your own control-don’t outsource that agency elsewhere.

Both of you are very academically inclined. Do you want your children to similarly excel?

Well, yes. While it’s early to comment on what their inclination might be, whatever profession they take up, I want them to give their best. We do come from a home where everyone is heavily invested in education. But it’s not just studies, I want them to take up different activities and do them well. This is important for their all-round education. They learn classical music and play cricket and they enjoy it.

Both, you and Om are musically inclined...

Yes, Om is learning to play the drums, whereas I am a classically trained singer and enjoy pursuing it.

A question for both of you. What are the qualities that attracted you to each other? Are you similar or different people?

Om (laughing): Well, after 20 years of togetherness that is a tough one to answer. Are we similar? I guess after all this time we are. But yes, Lavanya is a kind, friendly and really nice person.

Lavanya: He is very helpful and that’s a wonderful quality. Apart from this, I guess our interests are similar and we value similar things. He is very helpful and that’s a wonderful quality. Apart from this, I guess our interests are similar and we value similar things.

At the end of the day, what is your philosophy of life?

Lavanya: The one thing that comes to mind, quite simply is this: whatever be the struggle in life, the beauty and joy it brings through various moments of progress, must exceed the struggle.

BY KALYANI SARDESAI