Hi Nithin, great to meet you! Could you give us an introduction to yourself…
I studied my Bachelors in Physics which was nearly 10 years ago, thereafter I decided to do a MBA in Business Management. While pursuing my MBA, I was offered a position by Sainsburys as a Graduate Retail Manager. I worked there before moving onto Marks & Spencers. So nothing technical until that point!
Then about 2010/11 I was having thoughts like ‘this is not what I want to do’. I felt like I wasn’t using my full potential. And I work better when I am given a research task for example, which I can bring insights from. I wanted to go back to my science and physics roots. So I did some research into what I could use? At that point Data Science, was evolving, it certainly was not as big a field as it is now! I came across Data Science during my research and realized this is what I want to do.
So my story is different of course to someone who is taking the usual routes in data science, mine is certainly more of a career switch! The main thing for me, to achieve my goal, was to plan it. I didn’t want to tell myself: ‘I’m going to be a Data Scientist in 1 year.’ It was not realistic. So the main thing I realized was I needed to plan. So by the end of 2014 I wrote a plan: ‘I’m going to be a Data Scientist in 5 years’.
I decided to pursue a Masters degree in Big Data, from University of Greenwich. My plan was to spend my first year prepping myself for it, and then in 2016 I would start my degree. So during the first half of 2015 I learnt and practiced C/C++, then in the second half Python/R. I bought some text books for Linear Algebra, Statistics and Calculus, refreshing my basics in math.
I started my degree in 2016. All this time by the way I was working still at Marks & Spencer’s. My plan was to finish my degree in 2018, get into an internship, then become a junior data scientist my mid-2019 and finally a Senior Data Scientist by mid-2020. I am on track so far! I finish my degree this year, and I am already a Data Scientist Intern at the Rank Group which is an amazing company.
Was your original plan to pursue a career in Physics?
Yes I was always very interested in Science and Electronics since childhood. As a child I was always interested in watching series like Star Trek and loved reading science fiction. Star Trek showcased automatic doors, Mobile phones and tablets decades before they were invented and as a child I was so intrigued by it. I have always been a big fan of Carl Sagan. There was a series called Cosmos which was originally aired in the 80’s but was also re-telecasted in the 90’s. I used to watch it, and that is how I developed my interest in Physics.
I don’t know what made me do an MBA, after my graduation in Physics. I still don’t know. But the key for me was I realized that. I did something about it, and followed my passion.
Data Science is the field where I can put my passion for research, and science to better use.
How do you think your background in Physics has benefitted you in your Data Science role?
If you look at many people who are Data Scientists, you would see a lot of people with a background in Math, Statistics, or Physics. When you solve problems as a Data Scientist, the underlying key is Mathematics. Studying Physics involves a lot of derivations and solving numerical problems. I believe that it helped me strengthen my base in Math. So when I do a Regression problem now, and I look at the math behind it, it is quite easy for me to understand the concepts.
So is a big part of being a Data Scientist your ‘mind set’? A curiosity and a passion for problem solving?
Yes, absolutely. I believe the technical skills can be learnt. So it is all about setting yourself goals about what you want to learn. You never know what story the data is going to tell you. So it is all about keeping an open mind, being curious and not jumping into conclusions.
What was involved in the decision making process to become a Data Scientist?
I did a lot of research! Data Science was one of the options. Another option was an MSc in Physics. The other was to look into Nanotechnology or Electronics. Everything involved creating/building something new from what you have been given.
They were my options. I visited a lot of online communities and discussed it in these online communities. Perhaps, as it wasn’t clear exactly what a Data Scientist does, which is probably still the case, I had to put a lot of effort speaking to people on LinkedIn and in the communities to understand what they really do.
I was digging deeper to understand the difference between a Data Scientist and a Data Analyst. Who is a Data Engineer? I also spoke to people pursuing doctorates in Physics . Talking to them suggested a lot of people are going into Data Science after their doctorates. So when I asked them what would be my best choice, to become a Data Scientist was the overwhelming response I received. And actually my now Head of Analytics at Rank, has a Doctorate in Astrophysics. One of my other two Managers has pursued a Doctorate from UCL and other a Masters in Applied Statistics!
Meeting people from similar backgrounds only strengthens my conviction to become a Data Scientist.
Why do you think there has been such a rise in interest in Data Science?
I think it has a lot to do with the Harvard Business review suggesting it is the ‘Sexiest job of the 21st century’. I think the brought a lot of mainstream hype to Data Science. Recently I have been getting so many requests and messages on LinkedIn on how to get into Data Science. I always advise them to make sure this is exactly what they want to do. Don’t do it because of that claim, or because the money is good.
One has to have that passion we spoke about earlier.
What was the process like getting your first role within Data Science and what advice would you give others?
Based on my experience, selecting the degree and field of study is such an important starting point. Most Universities now offer specific Data Science or data related degrees which could be overwhelming, so do your research. It was actually my wife who came across this Masters degree in Big Data online and suggested me to look into it. We both went for the open day. After an intensive discussion with the professors on the course, its contents, and its scope and also with special focus on my background, I was quite satisfied with what they were offering.
So it is important to understand the offerings and the differences between the courses. A degree in Data Science would sound very vague to the industry specialist, as it is in itself a vast field. So don’t just do a ‘Data Science’ degree. Be specific. I am doing my degree in Big Data. Likewise, there are degrees specializing in Machine Learning as well.
When companies look at various profiles of the applicants, they usually look for specialized knowledge. Hence, being a specialist will give that slight edge in my opinion. For example, if someone with a General Data Science degree and someone with a specialization in Machine Learning applies for the same job, chances are that the specialist will get preference than the generalist.
The process for me personally was tough. Balancing full time work with my degree was hard. I would work until 3PM, then attend classes from 3PM to 6PM. After the classes, I would be off to the library until around 11 PM. So very long days. But you have to keep your eye on the goal!
I should also mention that Marks & Spencer’s as a company was very supportive about my dream.
Furthermore, while applying for internships, my advice to other aspiring data scientists would be not to apply for any Data Science role. Just as being specific in choosing the degree, be specific about what you apply for. When I was at the University, many companies had come and made presentations to our Big Data class. One among them was the Rank Group, who presented a project they were doing and advised interested students to apply so that they can screen them. Their project was to build a machine learning model to predict customer churn, which interested me.
Funnily enough I had built something similar before for one of my course works. I had made a prediction model, although at a smaller scale with around 10,000 rows of data in Excel. We were given the task that involved data warehousing, data mining, building a machine learning model and visualization of results. We were given the freedom to choose any data set. I sourced a customer churn data set, and built my model using p0 and R. The accuracy was around 80%, which I was happy with.
So as soon as Rank came in, it interested me. I had some experience in a similar project. I applied to only Rank. During the interview, they asked me why I wanted to be in this project, and what my understanding of the problem was. I could articulate my response properly because of my coursework. And whilst my sample size was small, it was a start. And they said from that I understand the concept, which was the important thing.
I must also add that our head of analytics who himself came to do the presentation at our uni during his busy schedule. He is a person who really want to help students to experience and learn the industry standards and it is really amazing to work with him.
So the message that I would pass on would be to be specific about what they are applying for. There are so many roles in Data Science. Look at your specialization and where you have experience and apply based on that.
Look at the kind of problems that could be of possible interests and what tools would be needed to solve those problems. Kaggle and GitHub can really help here. One Scrape/download data, play with different data sets and projects that interests you. Enter competitions in Kaggle. You don’t have to necessarily win, but at least if you have something to show, especially when you don’t have commercial experience. It would benefit at interviews and also help you tailor your CV. Also showcase your talent in github.
What advice would you give to someone on how to balance self-study and full time work?
For me, it all comes down to how much you want to be a Data Scientist. How much you want it.
I always carry my favorite book with me. It is called ‘Think and Grow Rich’ by Napoleon Hill.
It was written in 1937. The author studied successful people for 25 years and then derived 13 principles from it which he says is what is needed to achieve anything you want in life. One of them is Desire. In the book Hill says ‘to achieve anything you must have a burning desire’.
So when someone is messaging you asking to go out, or you get invited out for a meal, I always remind myself of the book. Regardless of how intense it may become, just as a small fire only gives a small amount of heat, weak desires give weak results. Strong burning desires results in strong big results.
It is also about visualizing oneself in the role. Imagination is a key factor. In 2015 I was nowhere near where I am now in terms of the skills. But I would visualize myself in the office, doing the job, which helped me believe it is going to happen.
What are the fundamentals skills for someone looking to become a successful Data Scientist?
The underlying theme for me is mathematics. Hence, it helps to have a basic understanding of Linear Algebra and Calculus. There are algorithms that could be used, for example to build a Linear regression Model in R, can call in the ‘lm()’ function. But if we don’t understand what it does, or the concept behind it, then that would only get us so far. So, it’s advisable to start with basic algebra and calculus to get a base in problem solving.
Then some statistics skills, like probability. Everything we are trying to predict is based around Statistics. So this teamed with the Algebra and Calculus will really round out the required Mathematical skills.
Programming skills could be learnt. I would suggest R or Python. Both are intensively used and both will do the job, but R was fundamentally developed for statisticians. Most companies in this field leave it onto the teams which of the two they would like to use.
Spark seems to be becoming more and more popular. It is written in Scala. Python and R can’t handle Big Data. Normally we would select a smaller sample and build and train the machine learning model in Python or R. After successfully training and re-training the model till we get a good accuracy, we would scale it up, and try it on Big Data using Spark in the Hadoop cluster. So learning Spark would also be a great addition to your skillset.
And finally SQL. It sounds obvious but to analyze the data, you need to retrieve it. So that’s the final basic skill I would suggest learning.
Once you start an internship or your first ever role, how important is the role of your first data science manager?
I am glad you asked that. It is extremely important. I cannot stress this enough. When I first started I was so overwhelmed by everything. The real life scenario was nothing like what I had learnt academically or practiced at home. I was introduced to so many tools and techniques. The enormity of data itself was intimidating. My experience before getting the internship was with may be 50,000 rows of excel sheet. But dealing with tables with 100 million rows was something else. But the Head of Data Science who I mentioned earlier has been amazingly supportive right from the start. He made sure that not only do I get used to these technologies and tools but also experience the business first hand by organizing tours to the casinos. He is very approachable and always gives me confidence that I am moving forward in the right direction which for me at this stage is extremely important. The whole team for that matter is extremely supportive and it plays a huge part in building my confidence. Rank is a great company with amazing work culture and I am really happy to be gaining experience from there.
During your learning process, you were very proactive in finding yourself a mentor. Is this something you would suggest others to find?
Yes, the impact is huge, the confidence that you get from it. Because the path is a lonely one. It is also a long one. You don’t know if you are going to get there, and how long it is going to take. The mentor is someone who is an expert, who has reached where you want to be, who can fill you with confidence on that journey.
I was fortunate enough to have Pete Williams as a mentor, and when I met with Pete he will always say ‘Of course’. It helps with confidence and direction. He was the head of analytics at marks and Spencer when I met him. He helped build my confidence not only technically but also for the application and interview processes. He always made himself available whenever I needed his advice. It helped increase my confidence and gave me a sense of direction. He gave me the confidence to contact potential work placements.
For example, he helped me secure an attachment day in the data science team in Marks & Spencer head office. I attended their daily meeting and met with the Lead Data Scientist which really helped with the visualization we discussed earlier.
The role of the mentors is really big. Mentors are there to support and to provide guidance. When you are looking for a mentor, I would also make sure I don’t waste his time. I was not calling him saying ‘how do I do this’. He is there to guide me. I didn’t just call him, I would always respect how busy he is, and schedule my calls. Mentors are someone you need to respect.
My final point is that there are no shortcuts. It won’t happen overnight. Mentors cannot help cut corners. Be patient, never stop learning, be thankful and enjoy the journey!