Keenan Crane
Frequently Asked Questions
Open Positions
Can I work with you this summer at CMU?
I won't be at CMU until Fall 2015, but please check back later!
Can I do an internship in your lab?
Generally speaking, CS research groups in the US don't have official "internship" positions, though we do sometimes support informal short-term visits (e.g., for a summer or a semester) that are not affiliated with a degree program, summer program etc. Most typically this kind of visit is associated with an existing collaboration between faculty at two different universities, i.e., faculty at one university might host a student of one of their collaborators. A better way to get the ball rolling with a particular researcher is to start a fun, interesting, engaging conversation about ideas (say, at a conference or via email). If this conversation evolves in to a concrete and compelling research question, there is suddenly a lot more incentive to host a visit.
If I have my own funding, can I visit your lab?
If you are self-funded, it becomes a bit easier to host a visit. But many of the comments above still apply: time is more valuable than money, and a good short-term research collaboration demands a compelling idea and a concrete plan of attack. When searching for a lab to host your visit, you can make life easier on yourself by proposing a specific project rather than simply asking "can I visit?" Doing so demonstrates initiative, and makes it easier to assess whether there are meaningful connections with the existing strengths and interests of the host research group. Still, don't be surprised if the answer is still "no!" Most people (myself included) are much more inclined to work with someone they already know personally—which is why it's a great idea to attend conferences/workshops!
I am an undergrad at CMU. Can I work with you over the summer?
Yes! Maybe! See here for information about summer programs: CMU Summer Research Programs. As discussed above, we would still need to find a good project. (Also note that I am not at CMU during summer 2015).
Can I join your lab as a PhD student?
Maybe! PhD admissions takes place once a year, with applications due in early December. See this page for more information: CMU Doctoral Admissions. Note that students are selected by an admissions committee; I cannot admit applicants directly. I can most easily advise students admitted to either the Computer Science Department or Robotics Institute—you should mark either or both on your application. Finally, please know that we get way more amazing applications each year than we can possibly accept. (While that may sound like a platitude you hear in a rejection letter, it's really true. You folks are amazing.)
What's the difference between the PhD programs in computer science and robotics?
You can find a description of the different programs here: CS Doctoral Programs. Very roughly speaking the answer is that there are different course requirements, but the research itself will depend primarily on your choice of advisor / research group.
How do I improve my chances of getting accepted to the PhD program?
Can I join your lab as a postdoc?
Maybe! Let's talk. Basically the answer is the same as for visiting students (above): since a postdoc is much shorter than a PhD, it's best if there is a project with some concrete goals and objectives, and that naturally builds on our mutual strengths and interests. Probably the most "real" way to end up doing a postdoc with anyone is to get to know them first (through conferences, etc.), and let a project naturally evolve. Having your own funding also helps. This list includes a small number of postdoctoral fellowship programs; if you're coming from outside the U.S., you should also check if your country has fellowships for studying abroad. Another possibility is to write a grant together. (Again, this all becomes easier if we already know each-other!)
How do you create the figures in your course notes?
Right now, the process for creating figures is really a labor of love. I start out in a 3D modeler (Luxology's modo), where I create the initial geometry, typically as a subdivision surface with a coarse control cage. I render this geometry out in several layers (silhouette, shadows, contour lines, etc.) and then trace over them by hand or using Image Trace in Illustrator. From there I choose the colorization, line styles, etc. For some of the figures I also write custom code (e.g., to construct a certain geometry). Someday I'd love to write a more automated tool, but there are only so many hours in the day! If you think you have a good way of attacking this problem, I'm all ears.
I am a recruiter. Can I interest you in a job at company X?
Being a professor at CMU is my dream job, and it's hard to imagine wanting to do anything else. That being said: if we're talking about seven figures and up, go ahead and shoot me an email! ;-)