how does grad school (for theoretical CS) work?

[malmadur]

I've never really known how some things in grad school happen/what happens in grad school, so I figured I'd ask here. (In case things are different between departments, I'll be a grad student in computer science this fall.)

1. What exactly does it mean when a professor gets a grant? For example, if a theoretical CS or math prof receives a grant for $500k over 3 years, what do they do with the money? I can see them maybe spending some on travel or faster computers, but I'm confused where all the rest goes. Does it mainly go towards funding students? (If so, what happens to professors with less funding -- do they simply have fewer students?)

2. If I've been guaranteed TA funding by my department, does this mean I don't really need to worry about who I want to work with (i.e., I don't need to worry if my professor can fund me), as long as I'm willing to TA?

3. What do CS graduate students do in the summer? I know many companies offer research internships (Microsoft, Google, etc.), but it seems like a lot of students stay at school anyways. Does summer funding usually turn up? (My acceptances only mention being funded for the 9 months during the school year.)

4. What is the relationship between you and your advisor? Is it mostly your advisor just suggesting a thesis problem for you to work on, and you meeting to discuss progress?

[belowthree]

I've never really known how some things in grad school happen/what happens in grad school, so I figured I'd ask here. (In case things are different between departments, I'll be a grad student in computer science this fall.)

I'm not really in theory, but I know the answers to some of your questions, so I figured I'd answer the parts I can.

1. What exactly does it mean when a professor gets a grant? For example, if a theoretical CS or math prof receives a grant for $500k over 3 years, what do they do with the money? I can see them maybe spending some on travel or faster computers, but I'm confused where all the rest goes. Does it mainly go towards funding students? (If so, what happens to professors with less funding -- do they simply have fewer students?)

A $500,000 grant over three years is a very typical NSF grant size. (All NSF proposals in the small category generally meet this. Typically the proposal asks for ridiculously close to 500k (like 499,938) and the NSF funds whatever they feel like of that. Sometimes it's significantly lower, but almost all grants propose at least $500k of funding.) This typically ends up funding about two students over three years. This may not seem a lot for 500k, but grant accounting can be tricky.

Typically 500k gets broken down something like this:

* A bit over half goes straight off the top to fund overhead costs. This part helps compensate the university for maintaining buildings, supplying power, that type of thing. So now you're looking at 250k. 250k over three years is about 80k a year.

* A bit of that usually goes to the professor for a month of summer funding. The more professors involved in the work, the more months of summer funding that might come out of the grant. (The NSF also limits the amount of total summer funding they'll give, so sometimes this component won't come out of a proposal at all if the professors are already at their summer funding max. This can lead to summer funding for students instead. ) A few spare thousand goes towards travel and equipment budgets each, along with whatever else your grant prep people can come up with. Let's say all this totals 20k/yr. (This estimate is high, usually for a 500k grant these type of annual costs are lower and more money goes to students.)

* So now we're at 60k/yr. Which means you've got about 30k per student: ~20k to cover stipends (usually lower), ~10k for fee remission (usually higher, sometimes considerably at private schools or out of state rates for public schools). (You never see this money, but the reason you don't pay fees is because the grant does.) Grad students tend to cost more than 30k/yr a lot of times, but this is about the lowest you can get a grad student to cost. Depending on how things are run at your institution, this part of the grant can be larger. (Again, this does not always translate into more money for you, it may just be your internal fee remission rates are higher.)

Whether or not they can shoehorn in summer support depends on how much the professor is making, whether or not they even want you around in the summer (some groups prefer their students get some outside perspectives through summer internships) and the internal costs of fee remission within the university campus you're attending.

A rule of thumb is for every 100k of grant money you can fund about one grad-student year of work in CS. (In medical or bio it's completely different because they have all kinds of crazy equipment costs.)

2. If I've been guaranteed TA funding by my department, does this mean I don't really need to worry about who I want to work with (i.e., I don't need to worry if my professor can fund me), as long as I'm willing to TA?

TA is a funding mechanism just like any other. You can use it as long as the department is willing to let you, however often departments prefer to fund only the newer students this way assuming students will eventually go and get real jobs on grants. That said, theory funding can be patchy and a good department provides TAships to senior students if needed usually.

Typically most departments prefer you not TA your entire stay in grad school. You should be doing research, so you should get funded as a researcher.

3. What do CS graduate students do in the summer? I know many companies offer research internships (Microsoft, Google, etc.), but it seems like a lot of students stay at school anyways. Does summer funding usually turn up? (My acceptances only mention being funded for the 9 months during the school year.)

If your acceptance only mentions 9 months of funding, then you may well only have 9 months of funding. Many many many CS students flock to silicon valley, local companies and a few other hotspots every summer to kick back and spend a few months of the summer joining a team in industry.

Summer funding from grants does happen. TAs sometimes even, though much more rarely. Usually summer funding only happens if there's extra money to go around. It can be a very fine line since most groups have a certain number of students they need to fund, it can be extremely hard to hit exactly the right amount of grant money, so most groups are either over or under. Department funds (here's where some of that grant overhead money comes back) can help backfill groups that are under (this is when senior students get TAs) while summer funding can be an excellent use of money when a group is lucky enough to have a bit more money than they needed. (Very easy to do if you get one extra grant accepted you didn't think was going to make it. You don't hear back for six months on these things, so you have to scatter-shot and if you end up being more successful than you expected, you might actually end up with more funding than students, this usually leads to getting more students, but it can also just result in summer funding or intra-department collaboration.)

4. What is the relationship between you and your advisor? Is it mostly your advisor just suggesting a thesis problem for you to work on, and you meeting to discuss progress?

Again, I'm not in theory, so this one I can't help too much in detail on, but I can at least tell you that the theory folks do the typical weekly meetings. Whether or not your advisor chooses the path of your research or whether you chose the path of your own work depends on the relationship you two establish. Eventually all students should change from being assigned work to choosing their own work as your career in graduate school progresses.

Edited March 6, 2010 by belowthree