The Computing Innovation Fellows Project: Strengthening the Field in Difficult Times

By Peter Lee, Ed Lazowska, and Anita Jones

In this difficult economic time many Ph.D. graduates would be lost to the research and education track if—due to severely reduced hiring by universities and research labs—they accepted positions that would not permit them to pursue their independent scholarly interests. Doing this would diminish dramatically the possibility of a future research career.

To address this situation, the Computing Innovation Fellows Project (http://cifellows.org) was conceived in February 2009 by CRA’s Computing Community Consortium (CCC). The National Science Foundation funded the project, and in just five months 60 new Ph.D. graduates in the computing field were designated to receive fellowships for scholarly appointments at the nation’s universities and industrial research labs. In doing so, the project aims to keep the fellowship winners in the research community “pipeline.”

Why did the CCC create this project? Our primary concern all along has been the strong possibility that we might lose the full benefit of the investment that the nation has made in educating these individuals—one can only imagine their potential research contributions. Furthermore, increasing enrollments and rising research budgets may soon create pent-up demand for hiring at universities and research labs. As the economy improves and adjustments to new budget realities are made, we can hope that many of these Fellows will be absorbed to fill the demand.

The CCC has the mission to foster the creation of visions for future computing research and to work with funding agencies to turn these visions into reality. In February 2009, members of the CCC envisioned the project and brought it to the attention of the National Science Foundation’s CISE Directorate. Realizing the need to act quickly, the CCC wrote a proposal and submitted it to NSF/CISE before mid-March. NSF/CISE, in turn, convened a review panel and responded in less than four weeks with a decision to provide funding for up to 100 Computing Innovation Fellows. Up to 60 CIFellows were authorized for the first year, with the remaining funding to be used for renewals and new awards in subsequent years.

Between March and July, the team worked closely and intensively to make sure that all of the necessary processes would be put in place as quickly and sensibly as possible. In parallel with the NSF review, members of the CCC Council and an ad hoc steering committee, led by Peter Lee, defined the fellowship application process, a selection methodology, and web resources support so that the process could be executed rapidly while retaining a discipline of merit evaluation. A selection committee to review submitted applications was also formed.

The CIFellows Project is in essence a “stimulus program.” It was designed to address the extraordinary economic situation as quickly as possible, and ideally in time to help this year’s new Ph.D. graduates. NSF on its own could not respond in a timely way. The existence of CCC and the track record that CCC has established gave NSF the confidence to move forward with this project, administered by CCC. When we created the CCC, no one envisioned an economic crisis and no one envisioned the CIFellows Project. Yet, only a community-sanctioned organization outside of government could have executed such a national-scale program on such an extraordinarily tight schedule.

The hallmark of this unusual project is that it is broad-based. This was achieved in three ways. First, awards were made to Fellow/mentor pairs: each candidate Fellow could specify between one and three potential mentors, each of whom submitted a letter describing specific mentoring plans for the candidate. The goal was to ensure a highly productive experience for both the Fellow and the mentor; the quality of these plans had a major influence on selection. Second, no more than two fellowship awardees could have earned their Ph.D. from the same university, and no more than two awardees could go to the same organization (university, industry, laboratory or not-for-profit organization). We had two goals in mind here: to ensure broad participation, and to build bridges by means of the CIFellows between diverse institutions. Third, we encouraged diversity of other forms—not only of institutions, but of research areas, individuals, and so on.

Interest in the CIFellows Project was extraordinary. In all, 526 valid applications were received. More than 1,300 established researchers registered as potential mentors. Evaluation of the applications was performed by a committee chaired by Peter Lee. The committee consisted of an ad hoc selection committee of 24 members plus a subset of the CCC Council members. In a show of remarkable responsiveness, all of the reviewing was performed in less than four weeks. Every reviewer completed all of his or her assigned reviews on time, and all 526 valid applications received multiple reviews. Final approval of awards was given by the steering committee with oversight by the CCC Council. The Fellows were selected on July 7, 2009 and the awardees were given preliminary notice on July 10. Final approval was given by NSF for these selections on July 29.

The 60 CIFellows awardees were educated at 43 distinct colleges and universities and will be mentored at 48 different organizations. Among the fellows, 40% are women; 11% are African American, Hispanic American, or Native American; and just over 70% are US citizens or permanent residents. Their scholarly interests span a wide range of subdisciplines, from core computer science research to computational science to computing education. In sum, this is a remarkably accomplished, promising, and diverse group. Fellows were paired with mentors who should be able guide them effectively.

Several years from now the community can evaluate whether the CIFellows Project met its several ambitious goals: to maintain an impressive group of researchers in the research and education “pipeline,” to build bridges between computing researchers and computing research organizations with differing characteristics, and by doing this, to contribute to the long-term intellectual vitality of the field. We are planning to hold a session on the project at the next CRA Conference at Snowbird.

Peter Lee (Carnegie Mellon University) and Anita Jones (University of Virginia) are CCC Council Members; Ed Lazowska (University of Washington) chairs the CCC Council.