Part 2 - Getting Started

By Mandy Korpusik '13

Applying to grad school is a long process that
requires a fair amount of work, but I 
didn't find it overwhelming because I had
huge support from my parents... and I started early. I decided to apply to ten
schools that had strong computer science departments, and had professors
working on interesting AI (artificial intelligence) and NLP (natural language
processing) research: Stanford, CMU (Carnegie Mellon), MIT, UW (University of
Washington in Seattle), Penn (University of Pennsylvania), Columbia, Cornell, UIUC
(University of Illinois-Urbana Champaign), Johns Hopkins, and UT Austin. I took
the GRE (Graduate Record Examination) a week after I arrived home for the
summer, which was just two days before I flew to New York for my DREU. As soon
as I returned home after the DREU, I began writing my personal statement and
the essays required for the NSF fellowship. By the time first semester senior
year rolled around, I had done most of the hard work. I just needed to actually
fill out and submit the applications. I kept a spreadsheet of all the schools
and fellowships with their deadlines, application requirements (i.e. number of
transcripts, GRE scores), and other notes (i.e. which professors already
submitted their letters, etc). 


Group at UConn

Taking the GRE

Jason Zhu wrote a great post last fall about preparing for the GRE
- check it out: /sites/default/files/blog_archives/pgp/2012/10/shooting-for-success-on-gres-by-jason-zhu-14.html.
Personally, I used a Princeton Review prep book to study for the GRE, and I
took as many practice tests as possible. I tried to cram in as much vocabulary as
possible because I felt that was one of my greatest weaknesses. The GRE is
different from other standardized tests I've taken because it's done on the
computer, and is taken in special testing centers. I wore headphones because I was
surrounded by people taking all kinds of other tests (some of which may involve
speaking sections), and they made me sign in and out, took my picture, and kept
my watch, phone, and food in a locker. The two half-hour writing sections come
first, which was a relief to get out of the way. However, the sections become
increasingly easy or difficult depending on your performance from the previous
section of that same type. I completely psyched myself out after taking my
first math section because I thought I did horribly and imagined that the math
sections kept getting easier and easier. In the end, though, I did the best on
math and not as well on writing and verbal. Columbia grad students reassured me
that math is the most important section for a computer science PhD anyway. Phew!


Writing the

Each grad school application requires a statement of
purpose.  I found that at this stage of
the process, reading examples and looking at resources was helpful. Professor Allen
Downey pointed me to this essay by a professor at CMU: (section 3.3 is useful). Also, Yiyang Li '11 wrote
a helpful blog on his application process:
/sites/default/files/blog_archives/pgp/2011/04/part-iii-of-the-grad-school-process-by-yiyang-li.html. I used his outline for my personal statement:  one paragraph describing my general research
interests, three paragraphs describing previous research experiences (which became
increasingly long as the research projects became more relevant to what I want
to do in grad school), one paragraph describing career goals, and one describing
why I'm interested in the research and papers of professors at that school. I
found that the professors you list in the statement of purpose are important, since
one of them will need to be willing to advise you in order for you to get
accepted into the school. 



In addition to the research statement of purpose,
most grad schools ask for a CV (curriculum vitae), which is like a long,
research-oriented resume. After looking at examples on Google, I mostly just
added my GRE scores, research interests, research and teaching experience, and
publications to my resume. Grad schools seem to like it if you have at least
one publication or conference presentation, so I wrote about my Computational
Modeling case study published in Professor Allen Downey's book Think Complexity, my poster presentation
at the Grace Hopper Conference, and a paper that I was planning to submit to a
journal at that time. I also highlighted my leadership role as SWE President.


Lab at Columbia. My
advisor, Julia, is standing in the back right.

Letters of Recommendation

Most of my applications required three
recommendation letters. However, my advisor Professor Julia Hirschberg
suggested that I ask for four recommendation letters because they would make my
application stronger (as long as all four were excellent), and in case one
professor accidentally did not submit a letter on time, I would have three
backups. I asked Professor Hirschberg, Professor Chang, and Professor Chandy to
write me recommendation letters because I did research with them. I also asked
Professor Allen Downey because I have taken seven classes with him (including
two semesters of working on his SCOPE team), and he had already written me a
letter for applying to the DREU. PLUS a little birdie told me that both his and
Professor Chandy's letters were great, so of course that influenced my decision
as well.

How do I Pay for
Grad School?  Fellowships!

I applied for four
fellowships: NSF (National Science
Foundation), GEM, NDSEG (National Defense Science and
Engineering Graduate Fellowship), and NPSC
(National Physical Science Consortium). US citizens often will apply for the
NSF and NDSEG fellowships. GEM targets minorities in science and engineering. Professor
Hirschberg advised me to apply for the NDSEG fellowship. The NSF fellowship
application is by far the most difficult application I submitted because it
requires three lengthy essays (a personal statement, previous research, and
research proposal). Since I wrote those essays before I wrote my statement, it
was relatively easy to use portions of the NSF essays to write my statement of

Posted in: