My Grad School Applications

I’ve applied to graduate school three times in my life­ — each time under rather different circumstances and with different goals.

1. The first time, I was applying to Ph.D. programs in Evolutionary Biology after earning my B.S. in Zoology. I wanted to be a professor. Pretty normal.
2. The second time, I was applying to M.S. programs in Computer Science after dropping out of the above Ph.D. program. I wanted to be a software engineer. Fairly abnormal.
3. The third (and hopefully final :D) time, I was applying to Ph.D. programs in Computer Science as a 30-year-old, having worked for a few years as a Data Scientist/Machine Learning Engineer. I just wanted to do research. Somewhat atypical, but not unheard of (I suspect I’ll be somewhere in the oldest quartile of Ph.D. recipients when I graduate).

Because graduate school admissions can be quite opaque (I still have no idea why I did or did not get into any of the programs I applied to), I’ve decided to share all the materials from my three application rounds and their results. I hope they can be enlightening and/or comforting to others who are currently going through (or have gone through) the admissions process. Another good resource for applicant profiles is thegradcafe.com. The applications are in reverse chronological order. My CV is here. You can use the dates to infer what would have been included on my CV at the time of each application.

Computer Science Ph.D. (2017)

Schools

UW received “over 2,100 applications” for “45–50 positions” (i.e., 2.4% of applicants would eventually enroll, although the acceptance rate was probably higher). Michigan’s admissions statistics can be found here.

I actually didn’t originally apply to Auburn… my enrollment there was the result of some truly remarkable luck. By the time April had rolled around, I had resigned myself to the fact that I wouldn’t be getting a Ph.D. in machine learning; however, my resignation was premature. Serendipitously, on April 12th, I saw this Facebook post by the Evolving AI Lab at the University of Wyoming. The Facebook post mentioned that a recent graduate of the lab, Anh Nguyen (who is behind some of the first work about fooling neural networks), was new faculty at Auburn (as of Fall 2017) and actively seeking Ph.D. students. I had been following the Evolving AI Lab lab for a few years at that point and was a big fan of their work (Wyoming was actually on my short list of grad schools, but the location wasn’t a good fit for me and my wife). Additionally, Auburn is my much beloved alma mater (not to mention the place where I spent the first few years of my life — my parents had me as undergrads), so it kinda felt like the post was written for me.

I immediately went to Anh’s website and noticed he didn’t have any application deadlines, so I emailed him that night. Anh set up a Skype interview with me the next week, and it ended with him (informally) offering me a spot in his lab (this was easily the fastest turnaround for any position I’ve ever applied to!). Had Facebook not bubbled that post to the top of my News Feed, I probably would’ve never heard about this opportunity. So, in some ways, I have machine learning to thank for this chance to get a Ph.D. in Machine Learning!

Age

30. This is an interesting perspective on being an older graduate student.

GPAs

GRE

Date: 2017–08–10

For the second time in my life, I was pretty disappointed with my math GRE score. I had taken four practice exams prior to the real thing and never scored below 167 (with one 170) on the math. For whatever reason, I seem to underperform at test time. With that being said, most of what you read on the internet says the GRE is just a filter for competitive programs and doesn’t actually influence their decisions past a certain threshold.

One interesting thing to note is that both my verbal and quantitative percentiles improved considerably from the first time I took the test in 2009 when I was in the 84th and 72nd percentiles, respectively. What happened? I think there are three possibilities:

1. I just had a better test day.
2. When the GRE underwent its major overhaul in 2011, it changed in ways that benefited me.
3. I got “smarter”.

Any of these factors could have played a role, but I suspect most of the difference was due to (2) and/or (3). It would be interesting to know whether or not all of the computer science, probability/statistics, and economics courses and textbooks I’ve gone through since Auburn were better preparation for the GRE than my undergraduate Zoology curriculum.

Letters of Recommendation

Will Benton (coworker), Erik Erlandson (coworker), and Vibhav Gogate (professor). One kind of fun 21st century aspect about my application is that I’ve never actually met Will or Erik in person.

Statement of Purpose

You can read it here.

Personal Statement

You can read it here.

Costs

Computer Science M.S. (2012)

Schools

April 8th, 2013 was a stressful day for me. At that point, I had already received three rejections (with another one likely), and I had still not received an official acceptance letter from UT Dallas (I had been unofficially accepted to UT Dallas over the phone in February, but I was beginning to wonder if it had been a mistake). In desperation, I decided to apply to a postbaccalaureate program at Oregon State (to which I would be accepted). Apparently, all it took was a little bit of panic to spur the universe on… I received my official acceptance email from UT Dallas the next day. You can read more about my transition from evolution biology Ph.D. dropout to data scientist here.

Age

25.

GPAs

Same as above minus UT Dallas.

GRE

General Date: 2009–04–11

Subject Date: 2012–10–05

As I mentioned before, I was pretty disappointed with my math score on my first encounter with the GRE. Taking the Computer Science Subject Test was probably a bad idea. I basically tried to teach myself a full undergraduate computer science curriculum in a few months (I literally had zero programming experience prior to the summer of 2012), and the test was actually discontinued six months later because barely anyone ever took it, so I suspect it only hurt my application.

Letters of Recommendation

Craig Guyer (undergrad thesis advisor), Callum Ross (master’s committee member), and Mark Westneat (master’s advisor).

Statement of Purpose/Personal Statement

You can read it here.

Evolutionary Biology Ph.D. (2009)

Schools

Age

22.

GPA

Same as above minus UT Dallas and UChicago.

GRE

General Results: same as above.

Subject Date: 2009–11–07

Letters of Recommendation

Robert Boyd (professor), Debbie Folkerts (professor), and Craig Guyer (undergrad thesis advisor).

Statement of Purpose/Personal Statement

You can read it here.