Applying to Grad School

Thinking of graduate school?

The first question to ask yourself is why grad school, which is related to the second question, which kind of grad school.

As far as CS grad programs, there are basically three kinds:

  • “Research-oriented” masters, including the Berkeley CS 5th year MS, where a key part of the program is to produce some publishable original research; “Professional” masters programs, which focus on advanced coursework, (usually) a capstone design project that does not necessarily represent original research, possibly combined with leadership/communication/business courses; and PhD programs.

Masters programs at Berkeley: 5th year MS, M.Eng.

For 5th year MS (only available to our own undergraduates), in practice, I set the bar for research to “something we’d feel comfortable submitting to a high quality conference or workshop” (even if it is not accepted). Others’ policies may vary. Berkeley CS takes very very few terminal MS students outside of the 5th year MS. This program is a great opportunity for EECS undergrads, but you will be expected to work on some research and you need a faculty champion who will agree at application time to be your advisor if you’re admitted.

Note that your SOP and faculty advisor letter are expected to clearly demonstrate (a) an existing research relationship that is already in progress and going well, and (b) a clear plan for what specifically will be accomplished during the 5yr MS “continuation” of the project.

General requirements for 5yr MS: >=10 units of grad-level EECS courses, >=8 units 100/200 level non-EECS courses, >=24 units total.

There is no hard GPA cutoff, but averages of >=3.7 are not uncommon.

“Professional” masters programs focus on advanced coursework, possibly a capstone project, and often a shared emphasis with leadership/entrepreneurship skills. The Berkeley M.Eng. program (7%-20% admit rate, comparable in competitiveness to PhD admissions!) is a good example: you take 8 units of leadership-related courses from the Fung Institute, 12 units of graduate-level EECS courses, and a 5-unit capstone/integration project, all accompanied by “concierge level” career services.

Recently, 8 MEng alumni were featured in Forbes “30 under 30” list, and 50+ companies have been founded by 10 graduates. See the Fung Institute web site for more stats and details.

Masters programs outside Berkeley

Stanford has a terminal MS as well, though I’m not sure if it emphasizes professional leadership as much as the M.Eng. does. Georgia Tech has an excellent and very affordable online MSCS, which leads to the same degree as their on-campus MSCS; it includes advanced courses with projects, but again, I’m not sure how much emphasis on professional leadership.

Other comparable professional Masters programs: Stanford, CMU, UCLA, Columbia, Michigan.

PhD (doctoral) programs

In a PhD program, you commit to spending the next 4-5 years of your life producing a significant piece of research (typically several papers’ worth) that advances the field in some fundamental way. This is an intense 5-year monastic experience that is nothing like being an undergrad. It’s more a calling than a thing you do. You really want to be sure that this interests you, as opposed to doing it “by default”.

I would be cautious about applying to PhD programs if you have no experience doing research: some people love it, some hate it, and if you don’t love it, it’s going to be a long 5 years since that is all you will do. The sooner you get involved in undergrad research the sooner you’ll know if this option is for you.

General advice on applying

Whichever you do, apply broadly, as grad school admissions are very competitive; “hot” fields like AI and data science may be harder to get into. Don’t limit yourself to the “usual subjects” list of top schools: talk to professors and grad students working in your areas of interest to find out which programs are strong in that area—you’ll be surprised.

You can always come back

You can leave school, go to industry for a while, and come back later. That’s what I did. That way you really know if it’s what you want to do. You can leave with a BS or MS and if you get good letters from an employer, you certainly won’t be in a weaker position to apply!

If you really love teaching

If you’re considering a teaching focused career, most college level teaching positions require (or strongly advise) a Masters but not usually a PhD; the exception is a few schools where Teaching Professors are expected to have PhD’s, Berkeley being one of the few notable examples. Teaching at the pre-college level benefits from a Masters but does not always require one, but may require other teaching practice credentials.

The 5th year MS is a great pathway to try out research (stepping stone to PhD applications), for industry, or to pursue a teaching position.

What matters for grad school applications in generall?

  • GPA is important but mostly to be above a “soft” threshold that’s probably in the 3.7 range.

  • Letters of reference are really important. Academic writers are preferred; many professors will get help from their PhD students who’ve worked directly with you when writing letters. Industry letters are OK, but make sure the writer knows how to write an academic recommendation letter – it’s not at all the same as a job reference! If they’re not sure, ask your research advisor/mentor/champion if they’re willing to have a talk with the letter writer to guide them.

  • Your research statement is important - not because anyone holds you to doing what you describe, but to determine that you understand the research landscape in your area and know how to write concretely about a plausible project.

  • Your personal statement is important because it contextualizes your achievements and (importantly) can give a sense of what else you can bring to your cohort of fellow students besides academic ability.

What matters for PhD applications?

The following is only lightly edited from the rubric used in Berkeley CS PhD admissions meetings each year:

“Look for strong PhD potential evidenced by, among other things:

  • glowing letter
  • existing research record relative to opportunities available to them
  • GPA / academic performance / chose tough courses
  • well articulated research interests
  • achieving a lot relative to context/opportunity
  • communication skills (essay, interview)
  • will be able to TA
  • contributor to community (lab, dept, broader community)”

Using LLMs to write your statements?

Don’t do it. Almost every strong PhD program at some point includes a personal interview with one or more faculty and/or current grad students, and it is obvious within 30 seconds if the person on Zoom is the same person that wrote the statements. It’s fine to have LLMs correct actual errors of style, grammar, and orthography, but the ability to communicate clearly, in your own voice, orally and in writing, is a key skill not only for PhD admissions but for your professional career. Any time spent honing communication skills is time well invested.

Pre-application PhD mentoring from the LEAP Alliance

The LEAP Alliance offers pre-application mentoring, facilitated by current graduate students at UC Berkeley or the University of Washington.