Tuesday, February 23, 2010

Guest Post: Update from the CRA Career Mentoring Workshop, Day II

(ed note: Jeff Phillips is at the CRA Career Mentoring Workshop. His Day 1 dispatch is here)

It is day two at the CRA Career Mentoring Workshop.

Today was all about funding, with speakers from NIH (Terry Yoo), DARPA (Peter Lee), Laboratory for Telecommunications Science (Mark Segal), and NSF (Jan Cuny). Jeanette Wing, the assistant director at NSF CISE also made an appearance at reception yesterday.

(ed. note: I just heard that Jeannette Wing is leaving CISE in July. This is sad news - she was a strong and dynamic presence at CISE)

NIH advertised having a lot of money (about $30 billion, compared to $7 billion in NSF). The NIH has many sub-institutes with many different topics, but all applications are funneled through grants.gov. Terry Yoo was very enthusiastic about us applying for a piece of his large pie. It seems a bit tricky, however, to fit a pure computer science project into one of these institutes, specific health-related applications are enough.

We all (CI Fellows) thanked Peter Lee who helped spearhead the CI Fellows program. He recently joined DARPA to head the Transformational Convergence Technology Office (TCTO or "tic-toe"), a new program that will oversee many funded computer science programs. See the DARPA_News twitter feed for information on DARPA solicitations. Among other goals of this office, is to eliminate harsh "go or no go" conditions associated with DARPA grants.
For young researchers, look for CSG or YSA programs, similar in some ways to NSF CAREER awards.

The Laboratory of Telecommunications Science is part of NSA. They hire many many Ph.D.s for advanced computer science research. He could not tell us specifics about what they do, but compared it to an industrial research labs (e.g. AT&T, Yahoo Research, etc.). Movement between research parts and non-research parts is more fluid and is pretty hands on. Even the theoretical computer scientists and mathematicians they hire often build systems to implement their work.

To get funding through them, they generally have close and specific collaborations with faculty. The best way to start a relationship is sending a student on a summer internship (maybe even an undergrad) or via a sabbatical.

Jan Cuny from NSF decided she did not need to convince us that we should apply to NSF; rather, she just assumed we would and gave a howto on applying for NSF grants. Most important tip: **talk to program officers!** (before you submit). Otherwise, it is hard to give specific summaries from her talk (the slides will eventually be online--definitely look for them). The presentation nicely demystified some of the reviewing process; such as how grants are reviewed and why she may choose a certain proposal (for diversity) that scored slightly lower than another unfunded proposal. The other key advice: follow the guidelines precisely and carefully, make it easy for reviewers, and focus the content section on proposed work, not existing work.

A parting thought. It has been great to see many friends who are recent faculty or postdocs, in areas who I might not meet in my normal conferences. But it was a bit odd to have such a large fraction of my competition for jobs in the next year or two in the same room. The funding agencies were definitely here advertising how to get their funding, but if we did not realize that this was important, we would probably not have much luck getting jobs. If you were a department looking to hire to someone, perhaps it would have made sense to come here to recruit postdocs to apply ? Although I guess that is a bit optimistic, as it is a hirer's market.

Is there some way that having many people looking for jobs all in one place can facilitate the hiring process, or has this been out-dated with the electronic age? I would argue that personal interaction is underrated, and would help universities figure out not just who has a great resume on paper, but is also great to personally interact with.


  1. On that last note - many other fields (incl./not limited to English, Sociology, and our closest relative, Math) have at least preliminary interviews at an annual conference (in December/January). Since more people get to have face time with department delegates than otherwise possible, I think it would benefit those applicants who aren't in the top 25% publication-wise but nevertheless would make excellent faculty. I can't imagine how that would happen in CS without a unified annual conference, but perhaps it could be distributed to the major topic conferences that happen in Dec/Jan (SODA, NIPS, ???). But that could be a lot of work for a recruitment committee that is hiring in multiple areas. Then again, for most postings I've seen there are usually only one or two areas plus an occasional "other areas will be considered". Moving to conference interviews need not eliminate the possibility of hiring someone who wasn't at a given conference.

  2. This workshop is an excellent example of a field gone bad. Suresh, I know that you know about the scientist surplus. We are training 10x as many PhDs as there are research positions. In the last year or two, the ratio has probably become much worse.

    Responsible professors need to recognize this, and help prepare their students for careers outside of research. However, here we have the CRA putting together a "career mentoring workshop" that totally ignores the reality. 90% of the attendees will not be able to get one of these jobs.

    I know they mean well, but think that the CRA is incredibly irresponsible and short-sighted. Somebody needs to call them on it.

  3. anon: While I understand your point, it seems only fair to point out that at three of the panels centered around the research process (planning your research career, mentoring students, and balancing work and family), the panelists were from BOTH academia and industry. So it's not like the other options are completely unrepresented.

  4. Anonymous:

    About 1/3 of the attendants were already faculty. And about a 1/3 were CI Fellows. Several of the CIFellows I spoke with already had academic jobs lined up for next fall which had been deferred for a year. I would be surprised if a decent percentage (>50%) of the rest of the CIFellows are not able to get academic jobs in the next year or two. Call me optimistic, but there are reasonable number of available academic jobs (see here: http://www.cra.org/ads/) even in a down year, and someone needs to fill them. Many of the best postdocs this year are CI Fellows, so you'd think they would compete for many of those jobs.

    (There were also a fair number of grad students there, and non-CIFellow postdocs. For the grad students, I would think you would have to be fairly confident in your future to go to such a meeting, and your advisor well-funded. I don't know if I would have necessarily recommended this meeting for graduate students unless you were on the job market this year and had a really really strong resume that you were sure you would get a job. There are other issues that you would need to worry about first.)

    Second, several of the speakers spent a few minutes to tame our expectations reminding us that we (postdocs and young faculty) had already achieved a lot (i.e. a Ph.D.) and not to be disappointed if we did not get a faculty job or tenure. Also as Suresh pointed out, there was ample representation from research labs.

    Finally, I realize that there are more Ph.D.s than research jobs (academic and major research lab). I don't think this should necessarily discourage people from getting Ph.D.s. I greatly enjoyed my time as a grad student and now a postdoc. I got 6+ years to work on open ended research projects of my choosing and get paid for it. Its been fantastic. Plus, I feel I contributed to the advancement of science in a meaningful way. How cool is that? How many people can say they got a chance to do that?

    I also feel strongly that if I do not manage to get a job in research (academic or major research lab) the possibilities for industrial jobs is now much broader than when I just graduated from undergrad. Sure maybe I could have worked my way up the ladder in 6+ years, but I would not have developed the critical thinking, analysis, and problem solving skills that I believe now make me much more attractive for interesting industrial jobs. More importantly, I feel these skills will quite easily translate as the industry job field changes, as its bound to do rapidly in these modern times.

    I would definitely encourage others to go to grad school. I think the key is finding an advisor who guides you (without too much restriction) to interesting problems.

  5. Update: slides for most talks are now posted here:

  6. @Annonymous and Jeff: To give a foreigner's perspective, while there is perhaps a surplus of academic applicants in America, doctorates are in great shortfall in much of the developing world, such that the number of faculty positions outstrips the qualified candidates.As such, the market value on these candidates is very high, and their standard of living excellent when one considers the salary/cost of living ration.

    But by the mantra that one needs quality institutions to do good research, isn't that destroying one's academic career? Well- while there is certainly a compromise, I don't believe it as pronounced as first appears. A Ph.D (and post-doctorate) builds up one's independent knowledge base and research skills, while establishing academic collaborators that one will have throughout one's career. At that point, an excellent institution certainly helps greatly, but what is more required is the time, space and financial independence a secure faculty position affords, and which is readily available to most developing world emigrants with American Ph.Ds. I'd be interested on Suresh's take however, on whether good research can indeed be done at a nascent research university or institute in developing world.
    I should note that it's possible that the developing world will also eventually face the same glut of Ph.Ds, albeit several years from now.


Disqus for The Geomblog