Applying and preparing for grad school in the computational sciences
February 17, 2015 2:17 AM   Subscribe

I'm heavily considering pursuing a PhD focused on statistical and computational approaches to scientific problems, mostly those in the biological or cognitive sciences. This is my third year of undergraduate studies, and while I have some research experience, it's not very substantial. Is it worth applying in my senior year when I'm not confident in the strength of my application, or should I wait another year? Also: what about industry internships vs. research experiences?

Context: I'm a computer science major (who has taken a number of math/stats major courses) at my state flagship university. This is the spring semester of my junior year. I've participated in a couple semesters (plus a summer) of research, but I didn't interact much with the supervising professors, nor was my work very substantial, amounting up to a few not-terribly-complex scripts. My grades are generally OK, with A-/As in CS/stats classes and Bs in math classes, but I don't know my professors much at all (partially because classes are big, and partially because I'm unfortunately a little too timid to attempt office hours).

Goal: I want to pursue a PhD in computational science (keeping options open), computational biology (more specific), or CS/statistics with a focus on scientific applications. Examples of interesting programs I'm looking at: Caltech, Georgia Tech, Stanford.

Obstacles: My research experience isn't great, and a related problem, I probably won't have enough strong letters of recommendation; it appears that PhD programs and fellowship applications generally require 3-4, while I foresee getting only 1-2 decent letters if I apply next semester (fall of senior year). This part of the application seems to be the driving force behind applying for PhD programs.

What I'm doing about them: I'm starting work in a new lab this semester, and it sounds pretty interesting. I think that if I work hard at this, it'll give me some valuable experience, and get on good footing with my supervising professor. Along with the prof whose project I'm finishing up right now, that would make two decent recommendations for applying for grad school in the fall.

Questions:
1) How strong of a recommendation can I really get from a semester's worth of working under a prof's grad student? Asked another way, should I apply to grad school next semester? If I apply this fall, it may not end up being worth the effort, stress, and time required for putting together the GRE, statements of purpose, and letters of recommendation, only to not get into programs that are a good fit for me. On the other hand, if I wait until my ninth semester (which I know I can stay for), I'll have another couple semesters' (plus a summer's) worth of experience and a better sense of what I'm doing. (It might get lonely after a lot of my friends graduate, though.)

2) Where is that other letter (or two?) going to come from? While I'd like to work in the aforementioned lab during the summer, they don't have the funding right now to keep me on as a paid assistant. Hence, I've applied to some summer research programs (think REUs) in relevant fields, and am also interviewing for some software engineering internships at tech companies. With any luck, I should be able to get at least an offer from a program in each of the two categories in the next month or so.

The question is that if I have two such offers, and both of them sound fairly interesting, which one would be more valuable to me? The research experience sounds like a good idea to bolster my CV for grad school and see what it's like to live and do research at another school, as well as giving me exposure to other folks going in a similar direction. Yet, a software engineering internship seems good for honing my coding chops (which may come in handy regardless of whether I end up in industry or academia), learning how to work with big systems in the "real world", and would pay (at least) twice as much. Is it common for CS/related PhD programs which are more theoretical/scientific in nature to accept recommendation letters from, say, engineering managers at tech companies? Even if it is, would it be that much less helpful than one from a research supervisor?
posted by sqrtofpi to Education (5 answers total) 2 users marked this as a favorite
 
Wait to apply to PhD programs until you are confident of the research you want to do -- ideally you should be at a point where there are specific PIs in the field that you are interested in working with on fairly specific things. This is for your benefit, so that you are a more focused student and applicant, but also because having reached that point will make you a stronger applicant.

Between college and graduate school, by all means get a well paying SWE job. Save as much money as you can, have fun, and in your spare time read the journals and figure out what you're really excited about. It might be too many things, so keep reading. It might turn out to be not enough interest after all, so maybe stay at your SWE job.

It doesn't sound like you're ready to jump into a PhD program, and as a CS major who is hirable as a SWE, you have the option of the best paying, most enjoyable "gap year(s)" available.

Regarding REUs, do one if it really fits what you're looking for in terms of topic, advising, and quality of life (e.g. location), otherwise at this point I would take an internship with a company you could work for after college. A consistent three semesters of research in the same lab under the same PI at your undergrad school is worth far more than a summer REU anyway, and with the position at your home institution you even have a shot in hell of producing some publications.

Good luck! You're really in an "embarrassment of riches" situation so don't get down about the little things. You're doing fine.
posted by telegraph at 3:34 AM on February 17, 2015 [2 favorites]


What's the rush? You will be well-served for a number of reasons to wait. I think as a general rule people shouldn't go into PhD programs straight out of undergrad.
posted by MisantropicPainforest at 7:27 AM on February 17, 2015 [3 favorites]


How strong of a recommendation can I really get from a semester's worth of working under a prof's grad student?

Depends on, in roughly this order: (1) what kind of letters the professor writes, (2) how good your work is and how hard you work, and (3) luck.

should I apply to grad school next semester?

Looks like you will have some research experience, good grades, and a real interest in the material. This is probably good enough to get into a good program, provided that your undergrad university has a decent reputation. My opinion comes from my current service on a computer science graduate admissions committee at a prominent research university.

I would apply in your circumstances. All other things being equal, it's worth it to get your training over as soon as possible. You can accomplish so much more when it's over.

If I apply this fall, it may not end up being worth the effort, stress, and time required for putting together the GRE, statements of purpose, and letters of recommendation, only to not get into programs that are a good fit for me.

If you want to go to grad school, you're going to need to do these things at some point. If you don't go where you want this year, it'll be easier to revise it next year than to come up with a new one from scratch. The result will probably be better too.

Don't apply to programs that you don't think are a good fit! This year, only apply to programs you think you might accept without hesitation. Obviously you can change your mind after visiting. But don't apply to "safety" schools. Your safety plan is to stay for the ninth semester.

Is it common for CS/related PhD programs which are more theoretical/scientific in nature to accept recommendation letters from, say, engineering managers at tech companies? Even if it is, would it be that much less helpful than one from a research supervisor?

It's unlikely the recommendation would not be accepted. I rarely find such letters useful, though. The process of research and the research environment is so different from what one goes through in private enterprise. Also I wouldn't assume that someone who had spent time programming in industry will code any better than someone who had spent the time in academic research.

That said, the internship may be more useful for you. If you think there is a chance you might not want to go into research this would be a good way to try something else out. If so, I would make my primary decision based on that rather than on what would burnish your application the most.
posted by grouse at 7:53 AM on February 17, 2015 [1 favorite]


I'm mainly familiar with PhD programs etc in biology, but ...I assume that it would be close and still applicable to what you plan to do.

Where is that other letter (or two?) going to come from?

Just adding this tidbit in case you haven't thought of it. Do you have courses (or did you you have courses) with small lab sections taught by a grad student/TA? Did they get to know you well and did you do well in the lab section and respective class?

So as a grad student I cowrote a few letters of recommendation with the prof teaching the course. The basic idea was that I wrote about the student from personal interactions/how did they lab, the prof of the class added the grad and stood by whatever I said. Both names went on the letter and endorsed the student. Anyway, if you think about it from that perspective, you might know a person or two; again, ask if it can be cowritten with faculty.

Re: REU. The only REUs that I saw undergrads participating in were at the same university - so the student would do research in their regular lab, and meet with the program (and do poster presentations, discuss science, blah blah blah). If this is similar to what you are describing, then it is a great opportunity for grad school training and the programs were typically small (so another person can write your letter) and you get to do intensive research in your lab over the summer.

I am also adding another possibility that you might not have thought about - you mention planning an extra semester and doing research. If you are not entirely certain as to what direction you want to go and it sounds like you can graduate (check around to make sure that these positions exist, that you would qualify, and they pay a reasonable rate for you), you could also graduate and do ~2 years as a research associate/assistantship (whatever they have to work in the labs) at your university or another university/med school/etc. The idea is that they pay you to do hands on work and you get to participate in the research. It is not uncommon to get papers and/or posters published and presented at meetings - you can also save up some money pregrad school (to supplement your grad school stipend).

Just some thoughts in case - it honestly sounds like you are in a good place in regards to grad school, whether it be a year from now or a few years from now.
posted by Wolfster at 8:54 AM on February 17, 2015


I would like to encourage you to think beyond your immediate goal of pursuing a PhD and figure out what comes after. A PhD is not a career, but rather an extended period of training that prepares you for a career as a researcher. Unfortunately, actual research positions for PhDs, i.e. faculty jobs at universities, are scarce and stupidly competitive; most never manage to enter into the career path they were supposedly trained for. Your odds of a good outcome are so much higher if you come in well-prepared, fully understanding what it is you are embarking on and what your goal is. Your ultimate goal, what the PhD is for.

The people I know who have gotten the most out of their grad careers have been driven by their interests almost compulsively. Often they are already practitioners or professionals and are already embedded in the professional community, sometimes with years of experience. They are there because they already have a career trajectory and this is where it has lead them -- they need the PhD to reach the next level. They blow through with a laser focus. Such people are not typical, but they seem to be the ones that end up with the faculty jobs at the end. Most grad students drift, they let their careers be defined by whatever grant their prof happens to have funding for, and at the end of the process they are confused and unsure about what to do next and in many cases end up less competitive then when they started out, having burned away some of their most productive years. Having seen a lot of this I really try to discourage people from doing PhDs unless I can see that they are very obviously being propelled in that direction by the force of their vision.

In your case the vision is not there yet. But that's okay. It's very normal, and I would even say most people who leap into grad school don't really have the vision, but just because it's normal, doesn't mean it's a good idea. This path is full of suffering. It can be an escape from the grind for a few years, but it has a fearsome grind of its own.

You're very fortunate to be in one of the only remaining industries that has jobs and viable careers. A stint in a big tech company is absolutely a good idea for you, building up coding chops is one of the best things you can do for yourself, and I would even say think about working full time for a few years and spend some of that time exploring the topics you're interested in. Such experience can only help. You may hear grad school calling to you, but listen to what it says. If it is offering an escape, a way to avoid the real world, a way to make you feel good about yourself, then it's a trap and will lead to suffering. Wait for a voice that says I know what I want to achieve and grad school is the obvious next step. The first voice will be vague, general, abstract, and surrounded by fear and anxiety, whereas the second will be confident, certain, and grounded in something specific and real.
posted by PercussivePaul at 12:04 PM on February 17, 2015 [2 favorites]


« Older How do you use mindfulness and meditation for...   |   Help me search Mythbusters by topic Newer »
This thread is closed to new comments.