My Completely Unofficial Advice Page
The comments below are purely those of Anant Sahai and do not reflect the official
positions of the University of
California, the EECS
Department at Berkeley, the other communications
faculty, or anyone else. The topics are compiled from the questions
that I get asked as an advisor and by people who are considering
coming to Berkeley. They may or may not be applicable to you.
For Potential Students
Berkeley is an incredible school. There is no limit to what you can
accomplish here if you are willing to work hard. The
faculty is great, the department cares deeply about both
undergraduate and graduate education, and your fellow students are
a wonderful resource.
For potential undergrads:
- There is very little (if any) grade inflation here, so if you do
well, you will have the satisfaction of having earned your marks.
- Some commentators complain about the use of graduate and undergraduate
student TAs in research universities. At Berkeley, you will find most
of your TAs (we call them GSIs and uGSIs) to be very bright and a nice
complement to the faculty. TA's are usually closer to you in age and
experience, and can often explain things to you in a different way
than the faculty can. In most courses, you will get access to both a
professor and a TA. This is a very good thing, not a liability. And
for undergrad students that get a chance to help TA a course, this
can be an amazing learning experience that doesn't just deepen your
understanding of the subject matter, but also improves a lot of the
"soft skills" (i.e. skills that exercise emotional and interpersonal dimensions) that are
essential to real-world success.
- Class sizes at all the major research universities can sometimes
get very large. But this is a bit misleading because the faculty are
complemented by a lot of other course staff like TAs, readers, even
volunteers. The current "Berkeley EECS way" actually tends to have
more individualized support available the bigger the class gets. This
can be thought of as a consequence of the law of large numbers.
- Research opportunities abound for students who want to
learn beyond the classroom. You have to take the initiative though, and there is much that could be done to improve this experience.
- Your professors are available for one-on-one
interaction. All you have to do is come to our office hours. If you
can't make the posted times, try to set something up by appointment.
- The most important thing for you as a student is to learn
how to learn. This is not obvious at all and it can be the source of
much struggle. You might have thought that the approaches that served
you well in high school are alright. Almost certainly, they are
not. You need to learn how to learn how to engage with subjects
deeply and at a fundamental level. Exams and homeworks are not
enough. Hard work is what is important, but you need to work hard in
a way that is effective.
For potential graduate students:
- The research culture here is collaborative and you will have the
opportunity, should you choose to avail yourself of it, to work
closely with many faculty members and fellow students.
- Research groups here vary greatly in size, but in the
theoretical areas they can be a lot more intimate (2-6 students
is normal for me, for example). The most important thing when choosing a graduate school is to make sure that
there are faculty there that you have good intellectual chemistry with, and that they have the
time and inclination to mentor you.
- The key to success is the quality of research that you can
do. It is possible to do stellar work here and to take real risks
knowing that the faculty will support you. Research can be an emotional
rollercoaster and support matters.
- In your research, focus on the fundamentals but don't be
unwilling to get your hands dirty with the sometimes grungy details
that you need to work out before exposing the truth.
For Current Students
- EECS is not a major for learning buzzwords or cocktail party
conversation, nor is it a matter of memorization. You need to
completely internalize the material so you can apply it. The best way
to learn a skill is to practice it. As such, you should understand
that the assigned homework and projects in your classes constitute a
lower bound on what you need to do. You should supplement your
assignments with drill problems chosen on your own or from your
- It is helpful and fun to form a study group in your classes and
to prepare for exams by trying to stump each other with questions
that you come up with on your own.
- (ugrads) Don't get stuck in "high-school mode" and
continue trying to game exams and skate by on minimal work. Take
appropriately challenging courses and do not let the CalSO orientation people
convince you to dumb down your schedule. You chose Berkeley for a
reason. If you are getting bored, move up to the more challenging
courses. At the same time, don't be afraid to ask for help. Everyone finds
something hard, and working hard with faculty and TAs to actually get it pays off.
- (grads) Don't get stuck in "undergrad mode." You are not
here to just take courses and do well on exams.
- (ugrads) Use Eta
Kappa Nu (HKN) liberally as a resource and join up if
invited. HKN can help you get the most out of EECS. Ask them
questions on all topics.
- (grads) Attend seminars regularly,
even those that are not exactly in your research area of
interest. Berkeley gets great talks and you never know where you will
get the key ideas needed to advance your own research in the
future. Two to four hours a week spent in seminars is a good
investment of your time.
- (grads) Attend the weekly graduate student social hours
and chat with your fellow students after seminars. Your peers are an
- (ugrads) Set yourself a goal of understanding some system
from top to bottom before you graduate. For example, you might want
to know how the entire process of downloading and listening to an MP3
or OGG file works. This can be a good tool for integrating knowledge
across different courses.
- (grads) Maintain a list of open problems that you are
interested in. Include everything on the list ranging from vague and
unformulated large problems (e.g. make internet work faster.) to
specific small problems that are already formulated.
- (ugrads) Take some upper-division mathematics or
physics courses while you are here. Don't become too narrow. For example, if you
are on the EE side of things, try to also take some of the CS 17x
series of courses. If you are on the CS side,
try to take some of the EE12x series. It is a joint department of
EECS for a reason.
- Get involved in research as soon as you can. It doesn't have to
be something big or impressive. Many of the faculty have small
projects or ideas that you can try out. Do not feel that you have to
stick with a certain project for your whole time here. This is a
research university with a host of projects. Take advantage of it. More important than the first problem
is the chemistry and mentorship that you receive. You are here to grow.