You could love it or hate it, but you cannot deny the fact that mathematics provides answers to some of the most pressing problems in the business world through various techniques – Data Science, Operations Research, Machine Learning, the list goes on and on.
Abhik Giri, our next pathbreaker, Operations Research Scientist at Bayer, solves different simulation problems in the crop science portfolio with the larger objective of addressing Bayer’s mission of optimizing agricultural resources.
Abhik talks to Shyam Krishnamurthy from The Interview Portal about working on optimization challenges in a diverse range of industries – media, manufacturing, supply chain, pharma, telecom, logistics, FMCG, by formulating mathematical problems for different business questions.
For students, if you are good in maths, do not opt for engineering, be a mathematician instead !
I am originally from Kolkata, where I did my schooling at Don Bosco School (1993-2005). My father worked in Allahabad Bank, and my mother was a homemaker. My hobbies include reading story books, playing cricket, table tennis, chess, and playing violin. I had a deep interest in Mathematics right from childhood – whenever I used to feel sleepy, I used to solve maths problems ☺
What did you do for graduation/post graduation?
I pursued MSc (Hons) Mathematics from BITS, Pilani (2005-09). This is an integrated 4-year course after 10+2.
What made you choose such an offbeat, unconventional and uncommon career?
While studying at BITS, I was not even aware that there are industry jobs in applied mathematics. I was planning to work for a year or so, then pursue PhD and move to academia/research. Because of my not-so-great CGPA and the 2008 global recession, I got placed at a small company in Bangalore in July 2009, where the work was of building dashboards and powerpoint presentations, not anywhere close to the work I was looking for. I got a limited flavor of Mathematics when i worked for an Epidemiology Forecasting project. I predicted the percentage of people who might be infected by different forms of cancer based on the prevalence of risk factors in that country (eg: percentage of population above 60, percentage of people who smoke, consume alcohol, red meat, are obese etc.)
The first break came when after 8 months, in March 2010, I got an offer from an Ahmedabad-based analytics firm – DecisionCraft Analytics where I saw the huge potential of applied mathematics in solving real-world problems. The project which I was handling was so interesting, that I decided to stay on this path.
How did you plan the steps to get into the career you wanted? Tell us about your career path
At DecisionCraft Analytics, I solved mathematical optimization problems for different clients, including India’s number 1 TV channel (where we maximized advertisement revenue at different revenue planning and sales realization points) and a leading Gujarat-based fertilizer company (this was a product mix optimization problem where we determined what to produce, how much to produce, when to produce in order to maximize net profit).
In December 2011, I moved to TCS R&D in Bangalore and thereafter I have been in Bangalore. Gradually as my network increased, I realized that Operations Research is the most niche area of data science and awareness about Operations Research is much less as compared to the more traditional areas of data science like Machine Learning / Artificial Intelligence. At TCS R&D, I worked in a network optimization project for a leading telecom giant where the objective was to meet as much demand as possible with the available capacity and keeping in mind other operational constraints.
After working at TCS R&D for 1.5 years, I moved to Genpact in July 2013. Here I worked on different advanced analytics projects involving predictive modelling, simulation, optimization for global top 5 pharma clients. Projects included price elasticity of demand, promotion optimization (pharma companies try to boost medicine sales by various methods such as sending emails to physicians, TV ads, sending sales reps to the physician to speak about the product, billboards in medicine shops etc., the model aims to find out which of these strategies is most effective), and how to optimally distribute incentives among physicians.
As I started getting little bored with pharma, I moved to Intel in November 2016, where I worked on inventory optimization problems for their supply chain. After a stint of 1.5 years at Intel, I moved to FICO (Fair Isaac Corporation) in May 2018 where I solved different optimization problems in logistics, supply chain, and manufacturing. I wanted to explore something outside optimization, and hence moved to Bayer in March 2021 to work on simulation problems in manufacturing. During my 11.5 years career in advanced analytics, I have had the privilege to work in different kinds of companies – ranging from startups to IT/consultancy to product-based companies. I have worked in many industries – media, manufacturing, supply chain, pharma, telecom, logistics, FMCG.
How did you get your first break?
I was working at a small company in Bangalore in 2010 when I realized that the work was really drab and I started applying to some companies in February. The first break came quickly when in March, I was interviewed by a small analytics boutique firm in Ahmedabad – DecisionCraft Analytics. After a couple of rounds, they offered me a job and I joined after 2 weeks. I was excited to know that some subjects which I had studied like Optimization, Operations Research, Graph Theory did not just exist in books, but they were very much used in the real world also!
What were some of the challenges you faced? How did you address them?
Challenge 1: Not knowing a particular programming language can cause hindrances in your career progression.
I learnt whenever I got time. Besides having depth in a few areas, it’s also important to have breadth.
Challenge 2: As I have worked across diverse industries, it took some time to understand the domain after joining a new organization.
I tried to learn as much about the domain as possible from other colleagues, business, documents, and the internet. Ultimately, we fit the maths into the domain and not vice versa.
Challenge 3: It is difficult to implement models into existing business processes because generally people are resistant to change and don’t want to take risks.
I tried to show some tangible benefits to the business and win their confidence.
Where do you work now? What problems do you solve?
I work as Senior Data Scientist in the crop science portfolio of Bayer.
We solve different simulation problems for the existing factories / yet-to-be-set-up factories.
What skills are needed for your role? How did you acquire the skills?
One needs to be able to mathematically formulate a business problem
Decent programming skills (the language you code in does not matter, nor is it essential that you need to remember the perfect syntax. What’s important is you should be able to think of the logic)
Ability to think out-of-box
Ability to present complex mathematical solutions in business terms
What’s a typical day like?
Modelling / coding / testing
Collaborating with engineers who are building the other front-end/back-end of the tool with the model as input/output
Discussing the problem / solution with business partners
Making presentations to demonstrate the output to the client/internal business partner
What is it you love about this job?
Ability to impact decisions and improve life with the work I do
How does your work benefit society?
Bayer’s mission is “Health for all, hunger for none”. And I do achieve that through my work. Our models ensure that we produce maximum crops in minimum land. This has a multiplier effect because minimizing land usage also ensures that carbon footprint is reduced as we reduce the resources spent on land (water etc.)
In my previous company, I was solving logistics problems where we were minimizing the distance travelled by vehicles. This, besides saving the company’s money, also reduces the fuel usage and contributes to a positive environmental change.
Tell us an example of a specific memorable work you did that is very close to you!
In my first company (DecisionCraft Analytics), we built a revenue optimization model for India’s largest television channel which was embedded in a tool we had built. I was asked to explain how to use the tool to ~200 users. As we were a very small company consisting of only ~40 employees, I was entirely managing this massive training exercise. This was the most challenging and exciting work I have done so far as there was immense resistance from users in using the tool and also due to the volume of users and their questions/feedback. It took a lot of time, effort and thinking to win their confidence.
Your advice to students based on your experience?
- Do not choose a career because someone else chose that path (your likes and dislikes, strengths and weaknesses might be completely different from his/her). Do not get discouraged by what your relatives/friends/neighbors said because you chose an unconventional path. Ultimately, it’s your life and you need to decide what you want to do. Please remember it’s not at all necessary to do engineering / MBA to get a good job.
- If you want to choose data science as your career, do not worry about jobs if you have the right skillset. There are plenty of great companies out there. In fact, the number of data science jobs today is many times more than it was even a few years back as a lot of US/Europe-based product companies are opening offshore offices in India. In addition, there are a lot of small companies and startups doing a lot of exciting work in analytics besides the traditional consultancy/IT firms. The opportunities will increase in the future as companies realize the value of analytics and competition increases, and companies want to make the most out of every dollar they spend.
Gradually move from an individual contributor role to a leadership role where I would work on few projects and mentor juniors to work on other projects. Instead of just executing projects, I would like to collaborate more with the business to understand how analytics can be applied in different areas.