Erudite Talks with Alumni-Article 5

Mihir Rana (2017-BTech-Mechanical), Master’s Student @ the New York University (NYU)

Introduction

• Tell us briefly about yourself.

I graduated from IIT-R in 2017 with a major in Mechanical Engineering and a minor in Computer Science. When I joined, I had but a faint idea of what I wanted to do with my life, not unlike most other people who were somewhat ill-informed as to what their discipline was really about, choosing only the “better” of a few familiar-sounding ones. The first two years were formative in nature wherein I mostly focused on rekindling old hobbies and discovering new ones, like playing football, watching series, and learning about programming/web development/machine learning. I was also part of Team KNOx (an automobile manufacturing group) until the third year, and after zeroing in on my future aspirations that same year, I co-founded the Data Science Group (DSG) in an effort to collaborate with like-minded peers on campus to learn/spread knowledge about data science.

• Tell us about the University you got enrolled in for further studies.

After my undergrad, I attended New York University (NYU) for a master’s in Data Science. It’s a 2-year program that offers courses centered around machine learning and artificial intelligence. The Center for Data Science (under which the program lies) has appointees from several other schools, like computer science, applied math, and statistics, psychology, finance, biology, etc., which translate into an extremely flexible curriculum offering courses from all these domains. Broadly, my research and studies were in natural language processing, computer vision, and reinforcement learning.

The decision to go for Further Studies

• When did you realize that this is what you wanted to do?

My journey was fairly exploratory in nature, and I honestly wasn’t sure about my decision to go for higher studies until I had actually been admitted into a few schools. By the end of third year, I had dabbled with algorithmic competitive programming, web development, and been reading religiously about ML for 2 years, along with having done an internship in the same. All that time, however, I was also an active member of KNOx and had been devoting a ton of time building our car for the competition, which of course also further cultivated my interest in the field. It wasn’t until I started appearing for internship interviews that year, and saw a very realistic possibility of getting through to a good core industrial/mechanical engineering company, that I had to choose between the two. I spent a few hours deliberating with my friends and family before my final interview and finally decided that mechanical engineering wasn’t my cup of tea. That’s when I grew sure of my career inclinations, stopped investing more time in KNOx, and fully dedicated my time to studying ML. Going for further studies (immediately), though, was still an open question. I was extremely late for the application process (and also missed a few deadlines) after being encouraged by the people around me to apply. It was when I had already started receiving positive feedback from a few colleges that I seriously evaluated my options and spoke to my brother and a few seniors pursuing further studies — all of whom advised me to “level up sooner rather than later” — that I made my final decision.

• Your motivation for giving higher studies more importance than securing a job through placements?

My situation was fortunate enough in that I didn’t actually have to make that choice in the beginning. I had secured a pre-placement offer from my internship last year, so I could devote my full attention toward applications, and if admitted, decide later if I wanted to go. That said, when the time came, for reasons I’ve mentioned below, NYU was an absolute no-brainer for me. In general, though, ML/AI is inherently a field where one probably won’t get too far without at least a master’s, so the real question for me wasn’t if I wanted to go, but if I wanted to go right away or after waiting a couple of years.

• Things you finally ended up doing to land yourself at a university for further studies?

I can’t say that I had actively been working towards applying for a master’s right from the start, but I can surely point out what helped and what I could’ve done differently.

First things first, GPA (fortunately or unfortunately) counts a lot, regardless of whether your undergrad discipline is aligned with the graduate one. It’s more about academic orientation than about knowledge in a particular field (not that that won’t help). I had also taken up foundational courses like statistics, linear algebra, optimization, and a minor in CS. Having a couple of internships in related areas and good ranks in online ML competitions added to them. As mentioned before, I also co-founded DSG in my final year, through which we studied, shared, and taught each other data science on campus.

In hindsight, internships abroad (especially in the same country as the one you’ll later apply to) tend to enormously lift your profile. The same goes for comprehensive research projects/publications, on top of taking up the bare minimum courses. (I didn’t have either.)

Standardised test scores do count, but not as much as people tend to think. 320/100+ on the GRE/TOEFL will open up pretty much all doors for you, so try to get this out of the way as soon as possible and focus on other aspects of your profile.

Most of all, your statement of purpose and potentially letters of recommendation will count. I won’t get into too much detail here, but *this is a blog I’ve previously written explaining the application process and how I feel one should approach it.

• Driving factors for the location and the University you chose?

Those who are familiar with AI and related domains know that NYU needs no introduction. It has been one of the frontrunners in this field. Some of the most groundbreaking research in AI has come out of its labs, with world-renowned pioneers like Yann LeCun, Rob Fergus, and Kyunghyun Cho teaching there, and I was extremely lucky to have been offered the opportunity to work with these entities and fortunate enough to be able to afford it. The MSDS program specifically was one of the very few that offered a full-fledged 2-year intense curriculum in such a nascent field, and also had strong research collaborations with top firms like Google/Facebook. Finally, of course, New York is one of the largest tech hubs in the world and has attracted big tech companies and startups alike, which was surely a bonus.

Experience

• The contrast between IIT Roorkee and your University?

I can unequivocally say that my experience at NYU was the best learning experience of my life. The flexibility in designing the curriculum ensured it was perfectly aligned with my interests and aimed at honing my skills. The major contrast between it and IIT-R — which, to be fair, can also be ascribed to the fact that this was a master’s program and inherently attracted more self-motivated students — was that of the culture and exposure provided. Most courses did not have any or <20% exam components and focused almost entirely on regular comprehensive assignments and projects. Exams were mostly open-book, and much like the assignments, were custom-designed each year, which precluded the success of rote learning. I could choose courses from pretty much any field I wanted, no questions asked. As mentioned before, most of our professors were also leading/working at top industrial labs, which guaranteed a constantly adapting curriculum and that there was no disconnect between the teachings and what actually “happened in the real world”, unlike what can be said about IIT-R to a great extent from my experience. We had company information sessions every Friday wherein their ML teams were invited to give a presentation and interact with us through the day, offering a direct opportunity for exposure as well as recruitment. Most importantly, for me personally, the opportunity to work with the aforementioned professors was something I could never have fathomed and is certainly my most prized experience.

Concluding Paragraphs

• Future plans

I’m very much still in the process of exploring my interests. In the near future (at least), I intend to stay in the tech industry and gain more real-world experience and find a narrower interest domain. Post that, I’m still unsure of whether I want to do a Ph.D or stay in the industry. At this moment, I feel the AI field is quite nascent and what I potentially might set out to do in a Ph.D may turn out to be a million miles away from what I actually end up doing — which is not necessarily a bad thing — but just something I want to be absolutely sure I’m prepared for.

• Advice to the campus Junta

My first advice (perhaps more applicable to myself) would be to develop an environment around yourself which incentivizes you to constantly explore new things. It took me a few tries to find my passion, and unless you’re one of the few lucky ones to already know what that is, exploration is the only way to realize it. Moreover, the process (and not just the final outcome) of elimination itself is paramount — it was an insightful journey for me that mitigated the future possibility of regretting not having fully explored something. Akin to the fallacy of sunk costs, it’s better to fail several times quickly (but not too quickly :) ) than to hold on to and pursue things you don’t believe in for an extended time, and college provides the ideal platform for it.

Secondly, know who you are and be confident in yourself. Don’t give in to the pressure — be it parental, peer, or something else. Don’t join clubs just because all your cool friends are; don’t fret over what others will say about you; don’t pursue a particular field because it offers better salaries right out of college; don’t be a spectator in your life. Take the wheel. Play the long game. Your notion of success will change in the future, and you absolutely need to safeguard it.

Good luck, and feel free to reach out to me at ranamihir@gmail.com!

LinkedIn ID:- linkedin.com/in/ranamihir

Personal Website link:- ranamihir.github.io

Link * :- https://medium.com/@ranamihir/how-to-apply-for-ms-phd-to-foreign-universities-b5d791eaa96d