Shweta Gupta, VP – Technology at Digital Vidya: This has to be one of my favourite interviews, very clearly. Here I talk to one of the most respected, acknowledged and profound leaders of the Indian Data Science space.
Kaushik has the experience of working at the intersection of technology, analytics and marketing globally for around 25 years. He has had leadership roles across shores & has built high-performance teams in such domains as Data Science, Modeling & Analytics, Business Intelligence, and Market Research. His areas of expertise ranges in providing Analytic and Big Data Solutions addressing Risk Management and Actuarial Pricing, Customer Experience Mapping, Lead Optimization, Customer Churn and Agent/Sales Force Optimization.
He holds a Doctorate in Marketing from the US and has published in international marketing and analytics conferences and journals including American Marketing Association, Academy of Marketing Science, and Decision Sciences Institute, among various other international journals.
Kaushik is presently the Chief Data Officer and Head of three Centers of Excellence (Business Intelligence & Research, Actuarial & Data Sciences, P&C Operations) at AXA Business Services. Prior to AXA BS, he was with Fidelity Investments in Bangalore where he was instrumental in setting up their Big Data and Data Science practice. He has also worked in such global organizations as Microsoft, Intel, IBM and HCL Infosystems, and has also had a stint in academia in the US. Kaushik is also active in analytics fraternity and is a frequent speaker at NASSCOM and other industry events.
I know that the introduction has got you glued to the screen. Let’s read what Dr. Kaushik Mitra has to share.
How did you get into Data Analytics? What interested you in learning Data Analytics?
Kaushik Mitra: My journey with Data Analytics had its seeds during my schooling days and it was fortified during my Graduate PhD years. I went to a school that was in the Guinness Book of World Records, and it was there that my love for mathematics started. The school had some really great teachers who helped nurture my passion for Calculus, Linear Algebra, etc. When I then enrolled into Graduate PhD Program in the US, I learnt various aspects of statistical and experimental designs, including predictive and prescriptive modelling techniques, etc. In fact, Structural Equations Modelling used to be among my favourite modelling approaches in Grad program. I still remember those days when, as Doctoral students, we would give a “run” on the system in the data lab, only to realize that the model isn’t converging – as a result we would end up sauntering to the lab in the middle of the night to change the parameters, and then again rushing to the lab in the morning to see the results. Those were the days! My relationship with this field got further strengthened over the course of my academic and organizational associations.
What was the first data set you remember working with? What did you do with it?
Kaushik Mitra: During my Graduate PhD program, I got to work a lot in areas touching upon Consumer Behaviour & Research, Advertising, Sales & Distribution, Services Marketing, Analytics, etc. The first large data-set that I worked on was provided by a professor of mine during my 2nd year of the Doctoral program. I still remember the efforts and struggle towards building a Discrete Hazard model to estimate time probability of an event. This work led to an article which got published in the American Journal. This was my first foray, and the journey continued thereon. My dissertation was in the area of Services Marketing where I ran multi-factor GLM models covering a wide variety of factors like primary surveys, behavioural and attitudinal data-points. This is among my earliest memories of working with large data sets. A generational leap has happened since then, and those data-sets probably wouldn’t be considered large enough compared to today’s standards. The fundamentals, nevertheless, remain the same, and that’s what, I believe, has stood me in good stead over the years.
How do you stay updated on the latest trends in Data Analytics? Which are the Data Analytics resources (i.e. blogs/websites/apps) you visit regularly?
Kaushik Mitra: I stay updated on the latest trends through multiple channels (both digital & non-digital). I follow with keen interest various online contents/blogs like Andrew NG’s contributions in Coursera; articles by Kunal Jain, Andrew Gelman, Gil Press, etc.; posts & write-ups in Simply Statistics, Dataconomy, Digital Vidya, etc; and forums like Data Science Central. However, my biggest source of learning and inspiration comes from my peers. I get to regularly interact with leaders from peer organizations as well as unicorns through the various conferences, data meet-ups, hackathons, etc. that I attend. I love the idea of interacting with them and learning through their real-life examples on how they’re solving important business problems through data science & machine learning.
Team, Skills and Tools
What are the different roles and skills within your data team?
Kaushik Mitra: I lead multiple Centres of Excellence across Business Intelligence and Research, Actuarial, and Data Science & Digital Analytics. The Data Science team includes both Data Scientists and Data Engineers, and they have capabilities in a wide array of areas of Machine Learning, such as Data Source Aggregation & Transformation, Feature Engineering, Data Lake Management, Analysis of Structured and Unstructured Data (Text, Voice, Image), etc. In order to keep up with the latest trends, the team is also extensively involved in developing use cases in AI such in Deep Learning & Neural Nets etc.
Help describe some examples of the kind of problems your team is solving in this year?
Kaushik Mitra: The Data Science & Data Engineering team has been involved in various exciting engagements within the Insurance and InsuTech domains. We have been working on problems in diverse functional areas ranging from Marketing, Process Engineering & Customer Experience ranging from Cross-Sell / Up-Sell & Churn Analysis to Claim Fraud Detection, Predictive Underwriting & Anti-Money Laundering. We are also working on AI, Machine Learning & Deep Learning approaches (e.g. Image Recognition / OCR, Natural Language Processing, Chatbots, Speech Recognition, etc.) for a variety of applications in our Process Automation journey.
Industry Readiness for Data Science
Which are the top 3 problems that are on top of Data Science, either based on industries or based on technology area?
Kaushik Mitra: Some of the key trends and developments that those working in the field of Data Science need to be cognizant and fully apprised of including the following, as these are shaping this field like never before:
- High-frequency transactions & distributed ledger technology, viz. Blockchain
- Alternate data sources, e.g. fetching data-points pertaining to small & medium-scale companies required for model building (which is otherwise not readily available using traditional means) through Web Scraping and Web Crawling techniques
- The increasing sophistication in Predictive Sequential Modelling using Deep Learning approaches (e.g., LSTM, RNN) for solving complex business problems
Industry Readiness for Big Data
Name 3 Industries and the kind of problems that they are solving using Big Data.
Kaushik Mitra: Some of the industries that are extensively investing in AI & Big Data include (but are not limited to) the following:
- Social Networking & e-Commerce (Google’s Search Engine & other products, Facebook Marketing, Amazon’s customer segmentation & targeting, etc. are prime examples)
- Retail & Investment Banking (using AI & Machine Learning to detect suspicious activity such as patterns suggestive of money laundering, increasing the odds of detecting fraudulent claims & transactions, etc.)
- Governmental agencies & military applications (detecting potential cyber-attacks, management of drones, etc.)
There are other industries, too, which have embarked on this journey and are at different stages of maturity. Organizations that have started investing into this space are better placed in creating/retaining their competitive advantages in the market, while others, who are still not bought into this, might end up facing the prospect of obsolescence in near and mid-term future.
Advice to Aspiring Data Scientists
According to you, what are the top skills, both technical and soft-skills that are needed for Data Analysts and Data Scientists?
Kaushik Mitra: Data Scientists & Analysts need to possess following 3 competencies and skill sets in order to differentiate themselves from the crowd:
- Excellent statistical and mathematical skills, including the sound understanding of various modelling techniques
- Proven coding and data engineering skills pertaining to Python, R, etc.
- Strong business/domain knowledge – given the rapid pace at which global businesses & industries are changing, learnability (or the ability to quickly un-learn and re-learn) is an essential associated skillset which needs to be possessed
What are the changing trends that you foresee in the field of Data Science and what do you recommend the current crop of data analysts do to keep pace?
Kaushik Mitra: Data Science will become increasingly commoditized in the future with the availability of powerful off-the-shelf products with comprehensive features and built-in libraries. Hence, in order for the current crop to be future-ready, they need to develop a deeper understanding of the science and fundamentals behind the various algorithms and data assumptions, and thus be able to gauge which model to use when & for which scenarios.
It is very important to not just possess strong Data Science & Analytics skillsets, but also deep domain knowledge. There exist tremendous opportunities to solve various existing operational problems & challenges (for which domain understanding is of paramount importance) – e.g. predictive maintenance of vehicles by advanced detection of anomalies through Machine Learning, AI-driven logistics optimization through real-time route optimization and behavioral coaching of drivers, enhanced customer service management through applications of Deep Learning and Social Media Monitoring, advanced claim fraud detections in insurance through NLP & Sentiment Analysis, etc.
The current crop would also do well to develop a specialized understanding of emerging data science applications (e.g. Edge Computing, IoT, Blockchain, etc.). The world is also rapidly embracing AI, and a lot of interesting things are happening in Deep Learning approaches, especially in the areas of voice, text and images. This technological / application-orientation can be a vital separator for those interested in creating a niche for themselves in the rapidly evolving techno-commercial space.
Would you like to share few words about the work we are doing at Digital Vidya in developing Data Analytics Talent for the industry?
I feel Digital Vidya is doing commendable work in driving awareness, knowledge and competencies in the exciting areas of Data Science, Digital Marketing, Analytics, etc. We need to be future-ready in order to remain relevant, and Digital Vidya is playing an important role towards ensuring that. Kudos to the team for driving this important agenda with passion and commitment!
To know more about Kaushik Mitra, you can check out his LinkedIn.
Are you inspired by the opportunity of Data Science? Start your journey by attending our upcoming orientation session on Data Science for Career & Business Growth. It’s online and Free :).