Exchange Diaries  

Atharva Naik | Computer Science and Engineering

Carnegie Mellon University, United States of America

Q1. While you were in college, what was your motivation to go for higher studies and in which year did you get that motivation? How did you manage your academics with your preparation?
A: I started going into research after joining the robotics research group because I was interested in delving deeper into the topics. However, I did not like robotics as much as I explored other fields as well. I started exploring the field of Natural Language Processing in my 4th semester by taking the NLP course and working on
the term project. I liked the experience of working on the term project which motivated me to do internships and
more projects to explore this field. As far as balancing academics goes, I tried to take courses that I was more
interested in, more in my research direction and that would benefit my potential applications as well. So I tried to
focus more on courses related to Machine Learning, Deep Learning, Reinforcement Learning and related topics. I
feel it's quite manageable if you take care of those things while selecting courses themselves.

Q2. What is the process of interdisciplinary studies? How easy was the switch? What motivated you to switch?
A: The department is Computer Science itself. While talking about MCU, it's not a department itself, it's more like a
tool. Within it, multiple departments are focusing on Robotics, Language Technologies, etc. My particular
programme is under the Language Technologies department in the school of Computer Science. So language
technology focuses on natural language processing, computational linguistics, etc.

Q3. What are the factors you considered for selecting the universities you will apply for?
A: I looked at the ranking of the universities, especially in the field of natural language processing as I was pretty
sure that I wanted to get into this particular topic. According to this CMU was the top institute. Apart from that,
other factors were that there are quite a few professors there, about 23 professors specialised in NLP - more than
most other institutes. It has a large research output and I also knew quite a few seniors and my sister who also
pursued her PhD from there. O I have a lot of personal accounts of the kind of research carried out over there. So
those were some of the factors that pivoted me towards MCU and applying for any other university.

Q4. Did you apply to multiple universities? If yes can you please elaborate on the differences in the selection
process of the universities? In universities in which you were not selected can you please share what you felt you
lacked, and what could you have done better?
A: I had applied for 10 programmes. Currently, I have been admitted to 2 and waitlisted in one. I got accepted into
MCU and the University of South California. I’m still waiting for a few programmes like those at Georgia Tech and
the University of Texas at Austin. I had also applied for a few PhD programmes as it was my initial aim but I
couldn't get in because of lacking research experience. So I applied to 3 PhD places and 7 Master's places.
For applying for PhDs, one has to have at least a specific amount of research. By research, I mean papers in top
conferences. So having a popular first author paper greatly increases one’s chances of getting into a good place.
Moreover, recommendation letters are quite important. I feel that while applying to American Universities it would
have been beneficial if I had a recommendation letter from a reputed American professor. In my case, I had one
recommendation from a Canadian professor and 3 recommendations from IIT Kharagpur professors. So overall I

feel that having a few first author papers and recommendation letters from reputed foreign professors
specialising in NLP might have helped me.
So if I’m talking about PhD applications specifically, the deadlines are a bit earlier. Suppose if applications for
PhDs were in early December, the applications for Masters would be mid-December, sometimes extending up to
February. Also, PhDs are more competitive than Master's programmes as there are fewer positions to fill. From a
materials perspective, there isn't too much of a difference. All universities expect you to have a Statement of
Purpose (SOP). Some programs may require diversity statements. Apart from this, the procedures were quite
similar.

Q5. How are you planning to manage your finances? Have you applied for any scholarships? What were their
procedures?
A: I haven't applied for scholarships. In some universities, you can get scholarships within the programme that
varies a lot from programme to programme. For example, Stanford University has a Knight Hennessy scholarship
program to which you can apply. But that has a very early deadline, like in October of the year which you applied
in, assuming the applications are in December. I didn’t apply for them. I was thinking about taking a loan or relying
on my sister as a sponsor. My particular programme doesn't come with funding but there are programmes within
MCU itself that are funded. Other than that, I plan to recover as much as I can from internships and being a TA.
That can handle most of your monthly expenses like rent, etc. But if you are quite focused on handling your
finances, you can apply to Canadian universities. There are quite a few research Masters programmes that come
with full funding.

Q6. Is CGPA a factor while considering higher studies at foreign universities?
A: From my experience CGPA was not a huge factor during applications. My CGPA was pretty good at around 9.6.
A good CGPA will certainly help but will not guarantee a seat. A CGPA of about 8.5 will not work against you but it
would be recommended to maintain a CGPA above 8 as usual. Also, it has more effect for Masters than PhD
applications.

Q7. Have you taken up any research project at your bachelor/master's level under any professor here at IIT KGP?
How did you manage your college studies and work for your research?
A: In IIT KGP my first project was the term project in my 4th semester on NLP. I decided to extend this project
while working with the same PhD student. It led to my first second author paper. After that, I did a bunch of
projects with external labs. For example, I remotely interned in the MAITC (Master of Arts in Technology and
Communication) programme at the University of Alberta. It was under a Professor who gave me a
recommendation as well. Another great way to do research is your DTP but keep in mind while applying that you
will not necessarily get a paper out of it unless you are already working with the professor and have enough
material to get a paper out of it to publish. Some of my friends were able to get a paper out before applying for
their Master's. You can also try working on research projects which could end up at the conferences I talked about
earlier.

Q8. What were the other difficulties you faced on taking this path and how to overcome them?
A: I did most of my internships in the summer so it was manageable. I did 2 internships together and I would not
recommend doing the same. But this was during the pandemic and they were online so it was doable. I would
suggest focusing on just one. If you are trying to get into a Master's programme, I would recommend putting
more effort into getting a research internship and working with a professor because that will get you much more
recommendations than doing an industrial. If you are not sure about what to do and you are in your 2nd year, I

would suggest getting a research internship and a CDC or industrial internship and comparing the experiences to
choose your path.
Initially, I was unsure about the job situation. Especially for many people, they're also sitting for internships during
the same time. In my case, I had got a PPO from my CDC internship but the PPO came quite a bit late. So there
were times when I was working on my applications while waiting for my PPO and there were a few other times
when I was working on 2-3 other projects while balancing the semester. So both of the issues involve multitasking
and time management. One thing I would recommend is setting realistic goals for your projects and setting
weekly goals. It would also help if you express your workload and be transparent with your project guides so that
they do not overburden you. Maybe you can change the timeline because research tends to be a lot more flexible.
But it is important to work on your academics too.

Q9. How did IIT KGP benefit you in your journey?
A: The main perk of IIT Kharagpur was the excellent peer group. I came across lots of really motivated people who
were working on multiple projects at the same time who in turn motivated me to do the same. Kgp also helped
me come in contact with researchers and internships using its wide network. Apart from that, the tag of an IIT
does help in the Master's application as well.

Q10. What are the opportunities after the PhD course?
A: After PhD, there are 2 main routes- one would be going into that would be applying for a postdoc and trying to
eventually land a position as an assistant professor. The other would be to pursue industrial research as a
research scientist or applied scientist. I am more inclined toward going into the industry. I would like to work on
projects that would have a direct effect on products. There are a lot of benefits to working as a professor too but
it's a lot more stressful. There can be a mixed approach as well. There are quite a few professors who work with
big companies while still holding their academic titles. So there are quite a few options to look into.