A career transition from Pure Science to Data Science is not as easy as it sounds, because such a transition requires a change in mindset from an academic, detail-oriented approach to an industrial, deadline-driven approach.

Debisree Ray, our next pathbreaker, Data Scientist, applies Machine Learning and Deep Learning methodologies to solve manufacturing problems in the industry by leveraging the power of data.

Debisree talks to Shyam Krishnamurthy from The Interview Portal about applying her strong research background in computational nuclear physics and numerical analysis, to the field of Data Science/Machine learning owing to the commonalities in the statistical analysis of large data.

For students, a successful career is built on the foundation of life skills, especially adaptability. Be focused on your goals, but remain flexible on the path !

Debisree, what can you tell us about yourself?

I was born and raised in a typical urban middle-class family in Kolkata, India. My dad worked as a branch manager in a central government bank whereas my mother, a science graduate by education, is a housewife. Growing up in an urban household, I never had to experience any kind of discrimination as a girl child. My parents always ensured that I was provided with the best of opportunities and resources that they could afford. They never pressured me academically, but were happy to see me grow up as a kind and knowledgeable human being.

I studied in Ramakrishna Sarada Mission Sister Nivedita Girls’ School, considered to be one of the best schools in Kolkata. Being a girls’ missionary school, it was not a rich and expensive school, but it had the reputation of imparting the best in terms of educational qualities. I must admit that this educational institution holds a very dear place in my heart. It has instilled a good dosage of life lessons, moral values, and work ethic in me which have shaped me into the person I am today. This is the place where I was encouraged to think independently and in a creative way. Apart from curriculum studies, I also took up acting and performed in the school drama club (though I was heavily influenced by my dad who performed group theatre during his college days). 

I must say that I was fortunate enough to grow up as a 90s kid in erstwhile Calcutta, where gifts like books (both fiction as well as non-fiction) and art supplies were very commonplace. This, accompanied by the creative environment provided in my school, goaded me to develop a great knack for reading in general. I started frequenting local libraries and had soon begun to grow my own collection of books. I started devouring books and gradually got a taste of both Bengali as well as World literature (mostly translated). This insatiable appetite for books is where I came across my favorite author Narayan Sanyal and his science fiction. His lucid yet informative style of scientific novels exposed me to the amazing world of astrophysics. The vastness and the intricacies of the known universe intrigued me. His novels had already struck a chord in my heart. This, added with my excellent understanding of Mathematics and Physical Sciences, had already encouraged me to forgo a career in either engineering or medicine. It was here that I decided to become an astrophysicist. 

What did you do for graduation/post-graduation?

I studied Physics during my undergraduate days at the prestigious University of Calcutta. It was here that I was exposed to the vast world of physics through some esteemed professors as well as getting access to study materials by foreign authors. I fared really well in my undergrad days and got admission for a master’s degree (merit-based admission) in one of the esteemed institutions, Rajabazar Science College (University College of Science and Technology, Calcutta university). I feel my inherent love for Physics influenced this vital decision in my career choice. I was really fortunate and privileged due to the fact that my parents never imposed any decisions on me and rather supported me whole-heartedly in every step of my career. I am eternally thankful to them that they have always encouraged me to follow my heart. With a master’s degree in Physics, a Ph.D. degree seemed to be the next logical step in my career. Hence, I started looking at Ph.D. opportunities both in India as well as abroad, which led to an admission to the Department of Physics and Astronomy at Mississippi State University, USA as a Ph.D. student. The opportunity to work in a reputed university like this as a Computational Nuclear Physicist really provided me the ultimate platform to pursue my quest for knowledge in the subject of Physics.

What made you choose such an offbeat, unconventional and unusual career?

There are a few people who influenced me to pursue a career in STEM, or more precisely, Physics !

I feel that I wouldn’t be where I am right now without the massive influence of Mr. Niren Brahma, my mathematics teacher in class 4. Although I was very young to understand anything about Physics or higher education, he was the person who envisioned me as a scientist and motivated me to take the path of higher education. So far, I think he was the most influential teacher in my life. I developed my undeniable love for mathematics through him. He employed unique techniques to explain the various concepts of arithmetic, and geometry, which made it really fun and intriguing at the same time to a kid like me. He simply decided to skip the traditional cookbook style teaching methods of Mathematics and piqued my interest in the subject. It is a testament to his teaching brilliance that he used to go beyond the confines of the syllabus and teach even higher level math even to 4th standard kids like us with absolute ease. I owe my undeniable love for Mathematics to him.

When I was in the 9th standard, an eminent theoretical astrophysicist, Dr. Sushan Konar (an alumnus of my school), came to deliver a lecture for the school students about her work and women in science. The topic of her lecture was physics of stellar bodies in the universe. That lecture was fascinating and right up my alley. I had already read Narayan Sanyal’s science novels on the universe. However, Dr. Konar’s seminar was particularly catered towards a younger audience like us. I listened to her spellbound as she talked and explained about different forces in the Universe. I was in complete awe of her as this was my first interaction with an actual scientist who worked in the highly interesting yet complicated field of astrophysics. It was probably on that day that this little girl had decided her future career path. I wanted to be like her. As they say, some facts are indeed stranger than fiction.

When I was studying Physics and Electronics during my undergrad days, I came into contact with a number of fascinating teachers. I have to mention at this point if it wasn’t clear already that I have always been influenced by unconventional teaching methods. It would be a huge mistake if I don’t mention Dr. Avijit Lahiry and Dr. Kamal Kumar Ghosh in this respect. Their unorthodox yet completely lucid ways of teaching coupled with their magnetic personalities heavily influenced me. Their methods of explaining both the basic and complex concepts were really instrumental in building a strong foundation for my career in physics.

Late Dr. Asimananda Goswami was one of the nuclear physicists (Saha Institute of Nuclear Physics), who used to teach Nuclear Physics during my post-graduation days. He was the first nuclear physicist who I had an opportunity to work for. He was a great mentor who both introduced as well as validated my love in the field of experimental nuclear astrophysics. I was so influenced by him that right after I finished my master’s degree, I joined his research group to work on a computational project with him for a few months. A lot of what I know today in the field came from working under him in those few months.

In 2010, I attended an international conference (SLENA, School cum Workshop on Low energy Nuclear Astrophysics). The week-long conference gave me exposure to various aspects of Nuclear Astrophysics. I got the golden opportunity to meet and interact with eminent scientists around the world. This has to be one of the most prized moments in my life, where I strengthened my will to obtain a Ph.D. degree in nuclear astrophysics.

I believe the turning point was the moment when I got admission (with full assistantship) to the Department of Physics and Astronomy in Mississippi State University, USA, as a grad student. I started my Ph.D. life in a different country, away from home. This provided me the ultimate stepping stone in my academic career. Moreover, this was the first time I was living thousands of miles away from my family and friends. I feel personally this was a drastic change for me. It’s one thing to pursue your career in a familiar environment but to do it in a foreign environment with absolutely new people and culture is a completely different ballgame. I feel these experiences coupled with the constant ups and downs of Ph.D. life have helped shape me as a completely different human being. I feel that this experience has actually taught me to take a step back and look at the bigger picture of life rather than visualizing through a tunnel. I have to say that I have learned some extremely valuable life lessons in this life-changing journey. 

How did you plan the steps to get into the career you wanted? Or how did you make a transition to a new career? Tell us about your career path

I started my higher education and Ph.D. journey as a nuclear physicist. My motto was to be open and flexible. Initially, I wanted to pursue experimental nuclear astrophysics; however, the department did not have any faculty at that point who was working in that realm. I joined the group of Dr. Anatoli Afanasjev to pursue Computational Nuclear Structure Physics for my doctoral studies, which led me to achieve a Ph.D. in 2017. My thought process and outlook evolved continuously. I initially wanted to be an Astrophysicist. But who knew that life planned something else for me? 

Working in the domain of computational nuclear physics is a two-way street. One is to unfold the mysteries of advanced nuclear physics. The other one deals with learning various aspects of computer science and engineering. I developed a strong background in scientific computing and numerical analysis in a high-performance parallel supercomputing (HPCC) environment. In addition to comprehending different nuclear physics concepts, my typical job description included developing computer codes (in Fortran language) and performing numerical data (large) manipulation, analysis, and visualization. By working on several projects, I gained deep experience in the statistical analysis of large data. With an already strong base in mathematics and statistics and an inherent knack for storytelling, I grew passionate about the relatively newer horizons of Data Science. During the final semester, I came across a Harvard Business Review article [1] that elucidated the new niche for physics doctoral graduates in the fascinating world of data science. This article piqued my interest in the subject. The opportunity to delve into the field came through my immediate post-doctoral position (2017-19) in an interdisciplinary engineering department. This is where I got my hands-on exposure to more real-world data-driven projects employing machine learning methodologies.  An additional online certification (2019) on ‘Data Science Career Track’ helped me expand my horizon in the data science realm more rigorously. I started to get my hands-on experiences in Python Data Science/ML Libraries. Working on capstone projects helped me apply all these techniques and introduced me to the Kaggle world of data [3].

As explicitly explained in the Forbes article [2], it is evident that ‘the Data is the new oil.’ With an increase in the usage of electronic gadgets worldwide, the data pull is enormous. The ease of availability of this massive data has ensured that major business decisions are highly reliant on these hidden data patterns. Nowadays, there is a huge inflow of PhDs (including me) into the industries from academia. Most enterprises’ data teams are employing Ph.D. Graduates with a background as academic researchers to deal with a complicated data-driven business problem. The real-world data scenario is noisy, unstructured, and complicated. Every data problem is different. Depending on the particular domain and the business problem, a data scientist needs to take a different approach. The data scientists are goal-driven versatile professionals who have to connect all the dots from data patterns to present the story to stakeholders; the main goal is helping enterprises with major decision-making and developing predicting models for future business directions. A data scientist needs to be a certain degree of an all-rounder. The typical job demands every bit of different aspects attached to it, such as the technical/engineering aspect, the scientific aspect (analytics), an aesthetic aspect (data visualization), and a storytelling aspect.  As a Ph.D., I was trained to tackle similar problems objectively, thus making this my dream job. 



[3] https://www.kaggle.com/competitions

How did you get your first break? 

I have already mentioned that my career goal changed towards the end of my Ph.D. career. There were many factors that influenced my decision. I was initiated into the newer horizons of Data science and machine learning. With my strong background in computational nuclear physics and numerical analysis, this change seemed the next logical step in my career. Additionally, I personally faced a real toxic environment in the lab settings, which bittered my intentions of working in an academic setting. (Many grad students/ research scholars often face these unfortunate work environments that affect their careers.) I wanted to focus more on getting an industry job. With this goal in mind, I got my first break when I joined as a Postdoctoral Scholar in an interdisciplinary Engineering department called ISER (Institute for Systems Engineering Research). I knew about the opportunity through LinkedIn networking, followed by a couple of rounds of interviews, including a formal presentation of my Ph.D. dissertation. This position really helped jumpstart my career outside the Physics Academia.

However, the ‘real’ first break in industrial settings came in 2020, when I was offered a Data Scientist position in Big River Steel LLC. I applied for this position through a reference in my network. Networking is the key to connect with people in any company. I went through a series of interview rounds before I got the final nod, including two remote interviews over the phone, tackling a project remotely with real-life data, followed by a final onsite visit and HR interview. Unfortunately, the Covid-19 pandemic started as I was about to join, and I was stuck with some visa issues. It took me a while to meander through those issues and eventually, I am excited to finally start my industry career in 2021. 

What were some of the challenges you faced? How did you address them?

Challenge 1: 

Challenges faced during Ph.D. life

The Ph.D. life is adorned with many challenges. After all, what is a rose without its thorns? My journey was also not devoid of it. First and foremost, Ph.D. research work involves taking up a novel project work that nobody has attempted before. So, technically it is a walk down unknown land. Graduate student life in a foreign land in itself is a life-changing experience and comes with its own challenges. A grad student has to survive this difficult phase through a minimal salary. Mere hard work and intelligence are not enough. It needs to be complemented perfectly with due diligence, persistence, patience, and of course lady luck herself. Rejection is a part and parcel of Ph.D. life. I feel that facing failure is one aspect that is rarely talked about. Failure to get publishable results often hampers the mental health of a grad student. Additionally, a toxic and hostile work environment (created due to bad mentorship), completely independent solitary living adds to the already prevalent difficulties of grad life. As an international grad student, I was also not immune to any of these difficulties.

The only way to cope through this problem is to keep calm. You have to take every opportunity you get to unwind and socialize in order to maintain rationality as well as sanity. Failure should be accepted as a basic part of grad life and shouldn’t be dwelled upon too much. You have to develop nerves of steel and look through the failure into the next step. You have to understand that this is a crucial learning phase in your life and by the end of it, you will develop the tools needed to survive anything. Time management and proper compartmentalization in order to keep a work-life balance is something I learned through this phase that helped me immensely to cope with these challenges.

Challenge 2:

Transitioning from academia to industry

Transitioning from the Physics academic environment to an industrial setting was another one of the real challenges I faced. Academic settings are more lenient, relaxed, and detail-oriented while industry jobs are more fast-paced and deadline-based. My strong academic background, familiarity with data analysis and visualization, coupled with my scientific computing skill sets were not enough to get a foot inside the industrial world. I had to align my skill sets according to the job requirements and had to take time to build my resume. The post-doctoral position prepared me for this transition where I learned new programming languages. I also updated my skill set by attending various boot camps, workshops, and online curricula on Data Science. I took a very cautious approach to observe and learn. I am still learning the industry-standard workplace ethics, and hopefully, soon will feel more confident about it.

Where do you work now? What problems do you solve?

As I mentioned earlier, I am about to start my position as a Data Scientist in Big River Steel LLC. Here I am being hired to solve Steel manufacturing business problems leveraging the power of data and employing Machine Learning and Deep Learning methodologies. Like every other industry, Steel industries are also upgrading to Industry 5.0. This means they are also employing the power of AI. Therefore, they are in the process of tackling their business problems in a much smarter way than before. To do this job, they are constantly hiring PhDs (in science and engineering) to tackle each data problem separately in a collaborative environment and explain the solution to the different stakeholders.

 The typical skill set required:

  • Computational coding (Python)
  • Data cleaning
  • Data visualization
  • Data storytelling
  • Exploratory Data Analysis (EDA)
  • Machine Learning/Deep Learning modeling (cutting-edge algorithms)
  • Report building, and presenting

What is a typical day like?

However, I can imagine, some portion of my day-to-day work will be dedicated to reading research papers, writing and implementing computer codes, data cleaning, model building, and presenting my result to the stakeholders.

What do you love about this job?

This is my dream job, as I get the opportunity to apply all my knowledge and skill set in detailed analytical work based on real-time data. Based on the said analysis, predictive modeling and vital administrative as well as manufacturer decisions, are taken. This is an agile workspace, where I need to update myself regularly with newer technologies. This fast-paced, challenging environment intrigues me, and I see tremendous growth opportunities as a data science professional.

How does your work benefit society? 

This is a two-fold question as I have to mention benefits regarding both my academic and industrial project works.

My academic research in the field of nuclear physics is related to Nuclear landscape and drip lines. As a nuclear physicist, my job involved studying the shape, size, and structure of the atomic nuclei. This is a fundamental question involving the basic laws of physics that will, in turn, answer important questions of celestial formations and the very origin of the known universe.  This research provides fundamental answers in our quest to know the macro (astronomical bodies) as well as the micro (atoms) world.  Knowledge of the atomic nuclei and their governing forces helps scientists develop vital knowledge, techniques, and research tools, which have a variety of practical applications, including energy production and exploration. Moreover, this kind of basic science research aids in the combined intellectual growth of humankind as a species.

My industrial projects involve working with the realm of data. With the surge of electronic devices around the world, an immense amount of data is available.  As an analyst/ Data Scientist our job is to find a hidden pattern in that data and use it for the welfare of the society or in the benefit of an enterprise. All these analyses will create ample opportunities for predictive modeling. This has far-reaching benefits from disease modeling and prediction to proper allocation and utilization of resources in an enterprise. 

Real-world data is a real puzzle. A smart and intelligent student will always feel intrigued by the challenges and complexities of a puzzle. Data is the next superpower. So, working with real-world data may provide an opportunity to change the world for good.

Tell us an example of a specific memorable work you did that is very close to you!

As far back as I can remember, I have always wanted to embark on a quest to pursue the very fundamental physics related to the birth and origin of the Universe. The first project work that I was involved in my Ph.D. tenure was to predict nuclear landscape and drip lines in covariant density functional theory.

As I mentioned earlier, one of the very basic questions that have always intrigued a physicist is the number of atomic nuclei in a universe. This has always been one of the most fundamental and open-ended questions related to the eternal quest of humanity to understand the birth and origin of the universe. To put this into perspective, each atomic nucleus in the universe must occupy a point in a graphical representation, termed as the nuclear landscape. This enclosure is bound by two boundaries known as the neutron and proton drip lines. The drip lines represent the limits of particle stability. So, the location of both drip lines measures the exact area of the nuclear landscape and eventually tries to answer the question perfectly. We undertook a systematic study of the location of two-proton and two-neutron drip lines (because our calculation is restricted to even-even nuclei) in the relativistic Hartree-Bogoliubov (RHB) framework with 4 different state-of-the-art covariant energy density functionals (CEDF)s. We also estimated theoretical uncertainties in the predictions of drip lines. Our results were then compared with those obtained in the framework of Skyrme density functional theory (SDFT) and existing experimental data.

Your advice to students based on your experience?

Most of the stories you hear have been those of success. You barely get to see the myriad of micro-failures that helped shape the success story as an impenetrable force. The key lies in not getting disappointed with these failures. Failures are a part and parcel of life. They are just mere bumps in this roller coaster ride called life. The real success lies in picking yourself up after falling down. Never get perturbed with the quantum fluctuation. Sometimes, it really helps to take a step back, and look at the bigger picture, and look at it with a clear mind. Learn from your past mistakes and move on.

Keep your spirit high and believe in yourself.  Self-belief and confidence are key factors for success. Analyze and evaluate yourself. Try to recognize your weakness and the strength. Professional and academic success is definitely important but life as a whole is more than that. It is really important to find a hobby or a good cause outside our professional life, that will provide you with the pleasures of living. This will help you keep your sanity and will play the perfect fiddle to your professional success.

Future Plans?

I am still in the early phase of my career. I wish to grow bigger as a Data Scientist and grab new opportunities and challenges as I get. Down the line, I can think of venturing to entrepreneurship by opening my own analytical farm. Who knows! One step at a time!