Speaker: Mike Walfish
Reccomended reading: http://www.eecs.harvard.edu/htk/phdadvice/
Point of a PhD is to teach you how to become a researcher.
Careers give you generally more intellectual freedom.
Bad reasons to do a PhD:
- Don't know what else to do
- Feel like you should do the next degree
- Prestige
Bad reasons to avoid a PhD:
- "I'm not smart enough" - you just have to be really good at your research area
- "It's a 5-year commitment" - some programs allow you to leave early with a Masters
- "Bad pay in academia" - in CS they're reasonable, because they need to compete with industry. Still a bit less, but not bad.
- "Solitary work is bad" - CS is pretty collaborative
If you have a strong undergrad in CS, you generally don't need to do a Masters.
Same answer to the question of "is this for me?": do research in a faculty member's lab, see if you like it.
Pick an area from the classes or talks you liked, then approach a faculty member who is active in that research area and talk to them.
Useful place to get an opportunity to do that is faculty research colloquia.
Recommendation letters are key, and project, work, and research experience should be reflected in them.
If you get your manager to write it, make sure your manager knows what a research position is looking for. Best to be a manager with an academic background, because they know what'll come across best.
Can also be a grad student you've worked with, who can recommend you to the prof and get that letter written from the prof.
Research statement is very important as well. Need to know why you want to do this.
GRE isn't all that important to Prof. Walfish, though some faculty members look at them.
TAing can help, if you can get it.
Don't rush the application. Take your time if you need to, even if it means applying later. You won't miss the boat.
Working in industry for a while won't ever be held against you and may be seen as a benefit.
Generally considered to be a bad idea to go to the same place for grad school as undergrad.
Once again, see http://www.eecs.harvard.edu/htk/phdadvice/.
Really, see it. Prof. Walfish read the whole thing aloud.
Multiple careers:
- Faculty member
- Work in a research lab
- Start a company
Running low on time here, so not much more here.
Many schools will often let you have some freedom of movement when you first come in, allowing you to look at the work done in many labs.
Masters programs generally don't let you accelerate your PhD, unfortunately.
Internships in the summer? Some, but not usually every summer.
A comparatively small amount of time will be reading papers—you'll be writing more often.
What does the word "research" even mean? Loosely: what a successful PhD student does for their thesis. More concretely: doing stuff no one's ever done before.
TAing while a PhD student? Same as internships: you should do a bit, but don't let it consume all your time.
Some universities require you to identify your advisors before applying. Contacting them is hard. Personal crafted emails are better than bulk or form ones.