Running grad student seminars

by Michael Ernst (mernst@cs.washington.edu)
written May, 1998 (and likely quite out of date by now)

The grad student seminar web page is https://www.cs.washington.edu/orgs/student-affairs/gsc/seminar/index.html; the local filename is /cse/www/orgs/student-affairs/gsc/seminar/index.html.

It's quite reasonable to reuse the previous year's list of topics and approximate schedule; this reduces your organizational effort considerably. At the beginning of the year, archive the previous year's WWW pages, such as putting them in a subdirectory like 9798/.

Publicity:

• Get the word to people early and often. Remind speakers (and attendees) a few times (a couple of weeks before, a week before, the day before, a few minutes before).
• Post to uw-cs.grads. Also consider grad-jobs, ugrads, ugrad-jobs. Don't post to uw-cs.talks (or talks@cs).
• Send mail to cs-grads-mail@cs nearer the date.
• Keep the grad student seminar web page up to date.

Scheduling:

• Schedule well in advance, so people know the event is coming up.
• Make sure the dates and times you set don't to conflict with quals-required courses by checking the UW CSE daily time schedules. First/second years are the most interested in many of these talks.

Room reservations:

• Reserve a room well in advance (3/4 weeks)
• Classrooms: Stephanie Pratt, stephp@cs (as of May 2000). She will coordinate with the university classroom scheduling system.
• Sieg 134: Scott Dakins, sjdakins@cs (as of May 2000). The department reserves this room every day from 2:30 onward and controls scheduling.

Refreshments:

• Ask Heather Dolan, hmdolan@cs (as of May 2000) for a half-order of seminar cookies, veggies, and soda. A couple of times a year, it's reasonable to get pizza (for instance, for gripe sessions and GSA elections where high attendance is important).
• Retrieve goodies from Heather beforehand and take them to the room.
• In general, serve food at the end of the talk. If you need to serve it early (for instance, if the speakers haven't yet shown up!), don't leave it out during the talk proper, as rude audience members will distractingly walk up to it during the session. Instead, bring it back out at the end.

Getting speakers:

• quals, generals: Frankye Jones, fjones@cs, can give you a list of recent quals/general passers; for quals, also ask for a list of those who have recently given their talks (even if they haven't yet officially passed quals).
• jobs: see the list of recent and imminent graduates, and also consider postdocs. Try to get a mix of people who interviewed/accepted jobs in academia, research, development, and startups. Or, have separate academic and industrial jobs seminars.
• It is usually more effective to ask likely panelists directly (instead of emailing or posting a call for speakers).
• When getting an outside speaker who visiting on a recruiting trip (like John Wilkes in November 1997), make sure that the affiliates coordinator (in 1997-8, Susan Eggers) is aware of the talk, and that the speaker is aware that the usual people are handling the recruiting aspects, and you are only coordinating the talk.
• Try to get people who have different perspectives on an issue to speak. E.g. for how to get a job, you can have faculty search committee people, people from industry, and grad students who have looked for jobs (both after a PhD and after a Master's). The former two groups may give their talks early in the year, so job-seekers for the same year can benefit from there. Ed Lazowska has given a talk along these lines early in the year for the last couple of years, but you can get people like John Wilkes, and other faculty search members in too.
• Ask the faculty graduate student advisor (in 1997-8, Carl Ebeling) for help in recruiting faculty members for talks. He's very helpful.
• You can crib from the mail I sent soliciting panelists for the Passing generals and Getting a job seminars.

Preparing speakers:

• Send out mail with suggested issues and questions (and encourage them to add topics you have left off); the presentation goes much better if the speakers have thought a bit about the issues ahead of time. Get topics from the publicity for the talk (possibly from a previous year). Here are the notes I sent to panelists to prepare them for various seminars:
  • Internships
  • Passing quals
  • Passing generals
  • Getting a job
• Ask the panelists to show up a few minutes early, so that the session can start on time (and the organizer doesn't get freaked out by their non-appearance, as happened to me).

Running the seminar:

• It's best for the moderator not to be one of the panelists (even if he or she knows something about the topic), because it's easier to moderate when wearing only that one hat.
• Introduce the panelists for their opening remarks. Your call whether to let questions be asked during this time — I think I would, unless they are really excessive.
• Add comments of your own, if the spirit moves you.
• Jump-start the questions with a few of your own, if others don't have any. The controversial ones are always good for a few exchanges.
• Keep the discussion on-track if necessary. In particular, watch out for people who want to talk all the time or who talk at length without saying much. You will find such people on your panels.

After the seminar:

• Return the leftover food/plates/cart; possibly leave some food out on the fourth floor of Sieg or other grad-student-accessible location.
• If speakers had overheads or other prepared materials, ask them for permission to include those on the WWW page.
• Consider typing up notes from seminars summarizing the discussion. This is a great service, though it's a fair amount of work. Michael Ernst's notes from the 1997-8 years are available at ../advice/. If you don't want to write up new notes of your own, you could just enhance those.
• Write a thank-you note (by email, probably) to all the participants. Be specific about what aspects of the event you were most happy about: that is far more sincere than a generic “Thank you for what you did”. If other students remarked to you that they found it useful, note that — it makes it seem that there is a groundswell of appreciation.
• Perform other followup, such as communicating feedback or proposing policies.

Suggestions for things to change or add

At the beginning of the autumn term, ask everyone in the department to write a paragraph about each internship they've had, to give people an idea of what different companies are like; then put this on a WWW page.

There was a suggestion that a few faculty that have spent time at research labs and academia have lunch with interested grad students who are making career decisions.

Gripe sessions should happen more often, say 2-3 times per year instead of once. And possibly two first year gripe sessions rather than just one. Gripe sessions should occur in the middle of a term: at the beginning no one remembers the pain of past problems, and at the end everyone is so grouchy that it's not really fair. It might be a good idea to have a gripes@cs email address that people can mail [anonymously?] to... anonymity is a powerful enabler, especially when it comes to gripes! The system administration staff should be asked to set this up.

You may want to work with job-seekers of this year to help arranging the “What's Been Happening While You Were Doing Your Thesis” seminar. Job-seekers have traditionally arranged it themselves and it is mainly for them, but others can also benefit.

Hold the “Jobs” seminar late. Those most interested in the seminar aren't taking classes any more, so the last week of the semester is fine, and fewer people are still interviewing or mulling over their job offers.

Back to Advice compiled by Michael Ernst.