Reflections one year into a PhD program

Reflections one year into a PhD program

5 Aug 2023

I started my PhD program about a year ago. In my first year I have:

  • Taken 4 “normal” 3-credit-hour classes
  • Participated in 3 seminars
  • Switched advisors
  • Attended 2 conferences (PLDI @ FCRC, JuliaCon)
  • Presented my work at JuliaCon

It’s been a lot of work, and there’s been a lot of stress. I’m in a much better place now than when I started, and over all I’m happy where I’m at and where I’m headed.

Changing advisors #

Some of the stress has come from finances: it’s hard to support a family on a PhD stipend with rising costs of living.

The most stress I felt was in my first semester. I got hired on as an RA1 with a nice researcher and we hit it off well. But I discovered half way through the semester that I was not enjoying the work. This wasn’t anyone’s fault—I had to do a little bit of the work to find out that I wasn’t that interested in what I was doing.

At first I thought that I might be able to power through. I didn’t want to be the kind of student that hops from advisor to advisor or from project to project just as the going gets tough, boring, or unpleasant. But I talked with my advisor from undergrad, and he explained that a PhD sets the tone of your research basically until you get tenure. That’s a long time—and I didn’t want to work in that area for a decade or better. I realized that I either had to change what I was working on, or I needed to quit and go back to industry.

I started looking around at what my options were. My school assigns every grad student a “faculty mentor”—a faculty member who is not your advisor that’s “assigned” to you. In my experience, most professors are more than happy to talk to you if you need help—this one just happened to be assigned to me. At some point I discovered that Ben Greenman was coming to the U as a professor. I called him, we talked, and we figured out that our interests were much more closely aligned. Problem was, Ben wasn’t starting at the U until the fall. I found a fellowship for one semester, and was able to do work with Ben and another professor through the summer.

Finding an advisor whose interests aligned with mine was a huge improvement for my work and my mental health. If you are looking into going to grad school, I would make that your top priority. And it wasn’t that my first advisor was bad in any way—we’re still friends, we chat when we run into each other in the hall, and I learned a lot from him—it’s just that our interests were less aligned than I originally thought, and it took a little time to discover that. Better sooner rather than later, though.

Reading papers #

One thing I noticed this week that made me really happy: reading research papers is a lot easier for me now. When I started getting interested in research back in my undergrad, I found reading papers to be so arduous. It took me a week or more to get through a single paper, and I never got a lot out of them. I felt like it was difficult for me to even understand the questions that the paper was trying to answer.

Now I’m more familiar with the context and the jargon, and I can grasp the questions the paper is trying to answer better. The format of research papers is familiar to me now, and that familiarity reduces the amount of friction I encounter when reading. It took time and exposure, and I don’t think there’s a substitute for that. Reading papers and discussing them with my advisors was a big help too.

The paper How to Read a Paper is probably the single most valuable bit of help an aspiring researcher can get. You should read it. (It’s short: 2 pages)

Classes #

Grad classes have a different tone than undergrad classes. As one of my professors put it, you will get an A or a B unless you persuade the professor that you should get a lower grade. This is nice because I have to keep a particular GPA to qualify for tuition benefit.2 That means I can do the minimum to learn the material and spend the majority of my time and energy on research.

After switching research projects, I noticed a switch in my priorities: my first semester, I would do my class work and then do research if I had time left over. My second semester, I did research first and class work happened when I had nothing else to do. I like the second balance better.

Conferences and speaking #

I got to present my work at JuliaCon 2023. I’ll put up a link to the video when it comes out, but if you poke around the recordings you should be able to find the talk in the middle of a long recording.

It was fun to prepare and present our work at a conference. It was also a ton of effort. I have never put that much time and energy into preparing a talk as I did for that conference. And it turned out pretty well, if I do say so myself. Hopefully subsequent talks will be less stressful, less effort, and even higher-quality!

JuliaCon was hosted in Boston; before that I went to PLDI in Orlando. Boston is far superior to Orlando. Orlando is a wasteland of hotels and conference centers and has nothing walkable anywhere. In contrast, Boston is a vibrant city that is a delight to walk around with plenty of interesting things to see.

It has been satisfying getting to know people in the PL community. It’s also been nice to work more closely with professors in my department. I’m lucky I get to hang around with so many curious, intelligent, and friendly people.

Satisfaction #

I’m happy with where I am. The PhD program was hard at first—and it’s still hard—but it’s a different, more interesting kind of hard, and I like that.

I think I’ll like the next few years. I still don’t know if I want to go into industry for a few years after or find a post-doc position in academia. Long-term I want to be a professor. We’ll see how it pans out though.

If you are looking for interns to study topics related to static analysis, macro systems, type systems, and language design, please drop me a line!

  1. Research Assistant. At least where I attend, there are Research Assistants, Graduate Assistants, and Teaching Assistants. RAs have funding and get paid to do research. GAs either have some kind of fellowship (some kind of grant/fund/pool of money from the school rather than a research grant) or are paying their own way. TAs teach classes in exchange for tuition and stipend money. ↩︎

  2. Tuition benefit: PhD students at my school in my field (CS) get their tuition paid for by the school. There are some obligations around this, such as taking a certain number of credit-hours each semester and maintaining a decent GPA. ↩︎