Now that I've become a faculty member, I realize that I must join the ranks of other professors that maintain on-line journals about their research . I hope to be at least somewhat as interesting as they are. I also believe that I am still under legal obligation to keep the content strictly about database research, so I'll do my best to maintain some decorum.
Job Search Reflections
A couple of people have asked for my take on this year's job search season. Others have already written extensively on how to be successful on the job market and I don't have much to add at this point. In particular, I found Matt Welsh's three-part series on How to Get a Faculty Job to be helpful. He was spot on in many ways about the entire process except for the part about wearing a suit and tie to the interviews. You definitely don't need to do that, but you do need to look presentable.
As for me, the best thing that I did during my grad student career was to follow what I call the "Dan Abadi Playbook." Dan was the lead grad student working on the C-Store project while he was at MIT. When Dan was on the job market, he was able to bill himself as the "C-Store Guy." This makes your research story during interviews much more compelling, as you can talk about the different parts of the system that you worked on. With a DBMS, there's always something for everybody.
The way I followed Dan's lead was to do the same thing but for H-Store. At the beginning of the project, there were four different PhD students that worked on the original system (two at Brown, two at MIT). But after a year, everyone else moved on to other projects. I then realized that I had the opportunity to become the "H-Store Guy." This was a hard road and I spent a lot of time building out the system; it took me almost two years before I was able to get H-Store running where I could then write research papers. During this time I setup the H-Store website as both a repository for me to document various things that I had added to the system, but to also advertise my work before I was able to get papers published.
I was also asked by multiple people during interviews about what software I used for my presentations. I only use PowerPoint through a VirtualBox VM. But I hired my undergrad roommate, Lauren Dellaquila, to fix up the graphic design of my slides. It made a huge difference and Lauren's help was money well spent. The change that she made to just my title slide alone is remarkable:
Tales from the Road
My interview season was about two months long. Everyone warned me that you can't get any real work done during this time and they were absolutely correct. It is grueling to fly into a city, get to the hotel, find a place to eat, and prepare for the next day's interviews. You can see how my productivity completed stopped during this period based on my Github commit record:
Here are some thoughts and highlights from my trips:
- The hardest interview questions came from Ippokratis Pandis and company at IBM Research. This was the first place that I interviewed at and I thought that all the other places were going to be like this. After I gave my one hour job talk, I sat in the middle of a circle and they took turns firing more questions at me. To my dismay (and slight relief), Mohan was not in town during my interview so I didn't have to answer questions on why H-Store doesn't need ARIES.
- Andrew Chien at the University of Chicago arranged for me to meet with researchers and scientists from Argonne National Labs who are working on some interesting projects.
- I enjoyed the dinner time conversation with Peter Dinda, my host at Northwestern. He's also a fan of the Hitch.
- Mike Cafarella and Prabal Dutta at the University of Michigan took me bowling. It is now memorialized on Mike's Wikipedia article.
- Margo Seltzer and Eddie Kohler at Harvard. Enough said.
- Rich Uhlig was a gracious host at Intel Labs, but since he was busy I ended up going to downtown Portland for dinner by myself. I unexpectedly ate at a restaurant that was half taco joint and half strip club.
- Lastly, I enjoyed interviewing at Microsoft Research and with the Microsoft CISL group. Microsoft has their interview process down to a science and were the only place that sent a gift basket to my hotel room. I greatly admire Phil Bernstein for his work on making transactions safe, but he ironically drove his Porsche through the streets of Redmond fast and loose.
And then of course there are my interivews at CMU. I interviewed here for three days and met a lot of amazing people. The one-on-one interviews at CMU were the most intense of all the places that I visited. I'll have more about to say about the people here over time, but the two major highlights of my trip are the extracirricular activities that Garth Gibson scheduled for me, such as the opening session of the Pittsburgh DataWorks consortium and visiting Google's offices in Pittsburgh, and the lack of cutting questions from Dave Andersen during my job talk. Christos Faloutsos and Todd Mowry were also so nice to me that it's almost suspicious. Needless to say, I'm quite pleased with the way that everything turned out.
New Faculty Hires
This seemed like the year that a lot of places were looking to hire a database or "Big Data" faculty member. The job postings usually couched it as looking for someone in "data intensive computing." Every database researcher that I know about on the market this year got an amazing position. We all interviewed at almost all the same schools. Here is where everyone ended up:
- Barzan Mozafari → University of Michigan
- Aditya Parameswaran → University of Illinois at Urbana-Champaign
- Spyros Blanas → Ohio State University
- Stratos Idreos → Harvard University
- Azza Abouzied → NYU Abu Dhabi
- Matei Zaharia → MIT
- Chris Ré → Stanford University
The last two don't really count, but I included them for completeness. Chris was already a professor at the University of Wisconsin-Madison. Matei is more a distributed systems person than a pure database researcher. I included both of them in the list because I admire their work and how they get things done.
If you're a grad student that is going on the job market this upcoming year, you may be thinking that all of the good schools have already hired somebody. Don't worry, because there are still lots of great places that did not pick up a new database professor and are likely to be looking again next year. The two most prominent are Columbia (with the formidable duo of Ken Ross and Luis Gravano) and UMD (with Amol Deshpande, my favorite DB researcher below the Mason-Dixon line). Other CS departments that I interviewed at that are likely to be looking again include Northwestern and the University of Chicago.
If you're on a faculty search committee and are worried that all the best people were already picked up, there are several students that I know that are going on the job market whose work excites me. Foremost, there is Jennie Duggan at MIT who works on SciDB, Alvin Cheung at MIT who works on computer-aided database program optimizations, and Aaron Elmore at UCSB who works on elastic transactional systems. And then of course there is Neil Conway (aka the "forbidden fruit") at Berkeley, who works on the BOOM Project. If you can wait a little longer, further down the grad student pipeline there is wonderkid Peter Bailis at Berkeley, who works on large-scale, front-end database systems, and Orestis Polychroniou at Columbia, who works on optimizing database systems on modern hardware.