This post is my personal story of why I decided to get a PhD doing research in graphics and vision.
Getting Into Graphics
While I was at Princeton, I took the undergrad computer graphics course in junior spring with Szymon Rusinkiewicz. I enjoyed the class, and I had a lot of fun doing the assignments. However, initially I didn’t think too much about it after the course was over. The following fall, while looking for fun CS courses to take, I noticed that the graduate graphics course fit neatly into my schedule. I reflected on the good experiences I had with the undergrad course, and decided mostly on a whim to take it.
Working on the various projects in the grad graphics course and rediscovering the things I enjoyed about graphics helped me to solidify the types of things I wanted to work on in a career. Working on raytracers and photon mapping and implementing spatial data structures highlighted the interesting algorithmic challenges that have attracted me ever since my high school olympiad days. We could implement viscerally satisfying applications like interactively deforming meshes, and we got plenty of concrete visual results from simulating complex fluid behavior.
The graphics group at Princeton was very friendly, welcoming me to many of the events and opportunities that were geared towards grad students. I was invited to the group retreat, attended the weekly graphics talks, and took a grad seminar on capture and fabrication. These opportunities showed me the extremely broad set of topics and applications that researchers in graphics work on, and gave me good a sense of what academic life was like.
To grad school or not to grad school?
Up to the end of my junior spring semester, I was quite certain that I wanted to go straight to Google after graduation. Somewhere around that point, I had a chat with a high school friend where we talked about augmented and virtual reality, sharing interesting videos and startups we had seen online. I got really excited about the topic, buying all sorts of hardware (such as a Kinect and an Oculus Rift DK1) and researching all sorts of topics (from hand tracking and gesture recognition to 3D user interface design issues in virtual environments), and experimenting with various libraries and SDKs. I took this excitement into the beginning of my senior year, where I decided that I wanted to make it my senior independent work topic.
My enthusiasm and willingness to work hard on 3D interaction projects even before the semester started made me reconsider my thoughts on what I wanted to do after graduation. I put together some grad school applications, with the initial idea that I would wait and see what options they would add to my employment offers from Google and Facebook.
Why Am I Doing A PhD?
Between my positive experiences with the Princeton graphics group and satisfying results from my independent work (a great demo, an A grade and a department award), I eventually decided that grad school was the right place for me. I really enjoyed experimenting with new technical ideas with viscerally satisfying, physical results. I figured that the best place to work on cool projects that could provide these results was in research, and I wanted to continue working on these sorts of applications while getting a PhD.
Further in the future, I would like to work in an industry research lab. Working in such a lab requires proven success in doing cutting-edge research, and this usually comes in the form of a PhD. However, besides getting experience and working on cool projects, I think that a more important part of being a grad student is making connections with people already working in industry research. The University of Washington graphics and vision group has very close connections with researchers from Microsoft, Adobe, and Google: I’m co-advised by Michael Cohen from Microsoft Research, and many other graphics students are also co-advised by industry affiliate researchers.
What about after grad school?
Large tech companies, such as Google, Microsoft, and Adobe, with the resources to support a dedicated research group, seem to me like the best place to achieve my goals. Because their core business is elsewhere, there is less pressure on the research arm of these companies to work on particular short-term tasks. Instead research can explore interesting projects that try to push the boundaries of what is possible with technology. These particular companies have the name, brand visibility, and resources to make people aware of any new technology and turn it into a polished product.
The idea of working in academia is actually appealing to me. I like teaching and I like doing open-ended research. However, there are two issues that make industry research more attractive to me. First of all, academic research is centered around publishing. Secondly, the supply of academic jobs is extremely low compared to the number of available PhDs. The competitive nature of both of these issues tends to reward more fundamental theoretical contributions instead of the applications that interest me.
Caveat: This is probably a gross generalization, and some people more into the startup sphere might want to argue otherwise. Too bad for you.
My opinion about startups is that there are too many of them. This means that any particular startup, even if it has good ideas, good execution, and a good product, will usually be lost in the sea of other companies and products. So if I want my research to have an impact, become famous, or “change the world”, the visibility and resources provided by large companies make them far better places to work. Once the right product idea comes along, and the bubble deflates, I may decide that a startup is the place to go.
So here I am, PhD student at the University of Washington in Computer Science and Engineering, and here I will be for at least a few more years. I’m here in the field of graphics and vision because I want to work on interesting projects with new technical ideas and concrete, physically intuitive applications. I hope to continue to work on these types of projects after I graduate, likely at a large company’s research division. Things may change some years down the road, but for now, this is the plan.