The content on this website is strictly the property of Insight and the Students’ Gymkhana IIT Bombay. If you wish to reproduce any content herein, please contact us:
Chief Editors: Shreerang Javadekar, Shreeyesh Menon
Mail to: firstname.lastname@example.org
Almost every summer blog starts with a recollection of how and when and in what environment they started writing the blog. Or perhaps they go on about why internships matter and the reason why they are so important to your future, ending again on the same note. Without further ado, I’ll get started because, well, you know, my code’s about to get compiled (or is it because I feel nostalgic all of a sudden?)
TransUnion LLC is an American credit information company, specialising in providing corporations and banks insights into the nature of their portfolios in terms of delinquencies, market shares and other relevant parameters. Currently, Transunion – CIBIL is India’s oldest and largest credit bureau and has access to credit portfolios of over 220 million individuals.
Besides American Express – which had its process in the first week of the internship season, this was one of the very few non-core companies that had openings for students in dual degree programmes in the first five to six weeks. After signing the IAF, it was a particularly long wait before we had any correspondence from their end. Finally, a test date was announced and the format was disclosed to us as a multiple choice round dealing with Java fundamentals, along with a couple of pseudo-code questions at the end. The interview short list came out soon enough and the people before me were, for the better part, grilled on the basis of their resume – their interactions lasting for half an hour on average. I wanted to secure this internship desperately primarily because it was one of the few companies that would allow me to play with their big data stacks and secondarily, who doesn’t want to get done with this process as soon as possible?
My interview, already delayed by over two hours, started with me discussing my penchant for cycling and gradually moved on to what, if anything, I knew about big data and Hadoop. I had spent the last couple of months learning the basics of what the Hadoop ecosystem consists of and practising variants of the fundamental algorithm behind most codes, called MapReduce. I was asked to write down the pseudo code for a simple MapReduce program and a few simple follow up questions on the same, following which the interview moved to an end with the obligatory discussion regarding the work they had to offer. I was out of the room in ten minutes.
Since the internship was to last ten weeks, we were asked to show up as early as soon as we were done with our endsems, which meant I barely had a day to myself after course project submissions. TransUnion has its office located in a very accessible location, hardly five minutes away from the Lower Parel station. We were the only two interns at the Mumbai office, both from IITB. We had a semi-formal interaction with the team on the first day, a couple of hours of interaction with the senior executives we would have to report to, followed by a lunch on the company’s tab. By the end of the day, the aim of the project was outlined to us and shown our desks and provided with a laptop. The first week was purely about getting a firm grasp on the tools to be used, namely Pig, Hive and Impala – yes, a lot of the products in the Hadoop ecosystem have names connected to animals. The practice exercises also included writing MapReduce programs in Java (Though we never really used it due to its huge boilerplate codes and mind boggling complexity).
Our work was related to facilitating the automation and transfer of programs written in SAS to Hadoop, while imparting complete configurability to the code. This project had two major benefits – firstly, the execution time was reduced from over 20 hours to barely an hour an half, secondly, the executioner need not fiddle with the code, having providing a configuration files stating his requirements.
This was the major project they expected us to deploy before we left, however, having completed the deployment in barely three weeks, we moved on to other projects including, but not limited to high-velocity batch processing in Spark and data visualisation in Tableau. The work consisted more of analytical thinking rather than the writing of long codes, thanks to Pig which scales down hundreds of lines of raw code twentyfold. Since we were tasked with performance tuning and speed tuning, we were done coding in the first week itself but it was not before we had survived over two weeks of testing debugging the code and tweaking the server configurations. It was a very fruitful experience playing with the Hadoop server configurations, learning how and why it actually works. TransUnion India has quite small a task force, as a result, our interaction was limited to the two senior consultants and the VP we reported to since we were the only ones who worked on Hadoop. As the placement folks came in, we were expected to guide them through the initial stages of learning so that they could take over once we had left.
The best part about the intern was that we could swipe in – swipe out whenever we wanted to, so our day usually began just after 11AM. Typical work hours would be around 6-7 during testing phases and floated around 11-12 during coding phases. We had access to an open pantry, with the option of having tea and coffee delivered to your desk. The area is in close proximity of multifarious good eating places besides being walking distance away from the High Street Phoenix mall.
In conclusion, it’s been a deeply enriching summer, getting to play with a stack as powerful as the one they had, to producing packaged code that actually helped the firm increase its revenue. Besides the coding, the visualisation exercises and the meetings with sales teams helped think like an analyst and understand like a salesperson. TransUnion has been a holistic experience, and has increased my liking towards this field of financial analytics to a level that I now consider this a decent exit strategy after my graduation.
A note of caution: The media has been throwing the term ‘big data’ around for quite some time now. I would advise against taking up an intern in the field if you abhor coding and are in it solely because you think it’s the next ‘cool thing’. A third year internship is usually the last chance to explore a plethora of fields before you graduate out of the institute and the safety net it provides. Make it count.