Back to The Software & Analytics Guide
My name is Kritin Agarwal, and I am a fourth-year chemical engineering undergraduate from Jaipur. This article is about my 8-week summer internship in the Analytics Profile in Piramal Finance!
Piramal Finance is an NBFC – non-banking financial company – which is very different from Piramal Pharma Solutions, a core chemical engineering company. In fact, they are recently demerged in the last week of August. It is also very different from typical finance companies such as retail banks – while PF gives out loans, it does not take deposits from customers.
This actually makes this role quite unique, with much to learn from the functioning of such a company. You also get to study a combination of analytics and finance, on financial datasets.
Why analytics?
At the start of the season, you apply everywhere. Knowing that I wouldn’t get a finance or core internship, my focus was more on software or analytics. Additionally, I am pursuing the DS minor offered by C-MiNDS, and did the Andrew Ng Machine Learning Course when COVID hit during my first year.
Selection Procedure:
After a resume shortlist, they hold a 2-hour long test. The test has 4 sections: DSA, English, Logical Reasoning, and Data Science, with 20-25 questions each. The next day, there were 9 candidates shortlisted for one interview round, of which 2 got selected. And just like that, my internship season was over in 15 days! During this time, however, it is very important to stay out of the ‘insti bubble’; talking to family helps, as well as having supportive friends and not comparing yourself to others, especially if most of your friend group is ‘machax’.
Coming to the interview preparation, know that this starts while you are building your resume. Every concise 3-4 word point is a lingering question that you must be thoroughly prepared with. You cannot start the day before. Additionally, have some company-specific questions ready to ask at the end of the interview (based on perhaps their plans for expansion). My interview was online, so make sure to do a preliminary tech check on your laptop to avoid any glitches.
My Time at Piramal Finance:
The internship, right from the interview, was completely online, much to my dismay. As most of my friends had offline interns, I also really wanted the same, and conveyed it to them but nothing could be done. Fortunately, in the last few weeks, they called back all employees, and giving the option to continue online or offline, my choice was obvious!
During the first week, there were introductions from different senior management about the company, its expansion plans, what they expect from us interns, and so on. I was in the Collection and Analytics Department (bank collections of those who default the loan), and many of my co-interns were allotted different departments such as Risk Analytics or Fraud Detection, so our work was considerably different.
For the next 4 weeks, my project was to build an ML model to predict whether the customer would default in the next month or not, using the available datasets of, for example, their payment history or sibling scores.
What I learned, first and foremost, was data cleaning. I feel like what we ignore about machine learning is that this data pre-processing takes up more than 70% of the job. There are many fancy ML algorithms that focus on while learning, but when working in an actual institution, the data you deal with is very messed up, to say the least. These fancy developments cannot be applied without many solely data-cleaning employees. That is not a very fancy job, but do not confuse it with a waste of skill. It is repetitive, but besides being an essential prerequisite for algorithms to work, it is interesting in its own ways. Data cleaning tells you the best thing a data scientist can have – that is the knowledge of data!
The rest of the timeline was basically end-to-end model deployment on AWS Sage Maker, enabling the model to trigger automatically instead of manually each month.
Looking Back:
This internship was a great blend of what knowledge I previously had from the courses I did or from other sources I learned, like DS303 and CL244. When you go to the job, they initially tell you to gain specific knowledge in topics like probability and statistics, numerical analysis, and the such for about two weeks. This ensures you have the technical knowledge to help when you code and apply models.
It was fairly easy to manage the online intern workload (any 5-6 hours throughout the day). Additionally, there was a whole hierarchy of mentors, of whom most were IITians. They are super friendly and very approachable and you can contact any of them. It was not a very competitive environment because everyone has their own task, but it was a good environment to work in. I had my mentor, a vice president, and under him, an assistant vice president, a senior manager, and a manager, and they were fantastic!
I wouldn’t say it was something very transformative, skills-wise. However, I learned to be as approachable and inquisitive as possible. Connect with and learn from everyone, and back in the institute, apply this to your professors and seniors as well. More than just my allotted mentors, I got a flavor of the true machinations of the company by interacting with people across all the different departments and walks of their careers.
I would definitely recommend such an experience and I wish you all the best with your internship season! :D