EECS lecturer Jason Ku ’09, PhD '16
By Grace Song
EECS lecturer Jason Ku ’09, PhD ’16 is leaving MIT after nearly 15 years at the Institute. Ku will join the National University of Singapore (NUS) starting next term, where he will teach mechanical engineering classes and conduct research.
For the past three years, Ku has taught 6.006 (Introduction to Algorithms), which in its most recent offering had 289 students. Ku is also known for his origami designs and computational geometry research.
The Tech spoke with Ku over a Zoom call to discuss his experiences as an MIT student, researcher, and lecturer. This interview has been lightly edited for length and clarity.
The Tech: Having been an undergraduate student, graduate student, postdoc, and lecturer at MIT, what drew you to come to MIT and stay for so long?
Jason Ku: I didn’t originally want to come to MIT. My dad is a Harvard alum and was in the Department of Engineering and Applied Sciences at Harvard, so throughout my childhood I was thinking I’d go to Harvard. I applied early to Harvard and got deferred. Then, I actually had to think about where I wanted to go. I ended up getting into MIT and was choosing between various places, but when I came to CPW, I definitely knew that I wanted to come to MIT.
I was an orientation leader almost every year when I was an undergrad. I told my orientees that the thing I loved most about MIT was that everyone comes usually not being at the top of a social food chain, so it’s not as cliquey as other places and people are open to making new connections. I came, as most MIT freshmen do, as kind of a socially awkward nerd. It was important for me to get socialized and learn how to enjoy being with other people, because it wasn’t something that came naturally to me. That was the most valuable thing from my undergraduate experience — more than my academics.
TT: How has MIT and its students changed over the years?
Ku: That’s hard to say. A lot has changed with my interaction with MIT and the MIT community, and I’m not sure whether I have a good perspective on how a particular subset of the MIT community has changed over time.
Course 6 has been the most popular major at MIT, but that definitely expanded during my time. When I joined, Course 2 and Course 6 were basically on par with each other, but now Course 6 is much bigger. So that’s been a change.
There are aspects of the undergraduate community that are an important part now but were not at all a part of my undergraduate experience, like MIT Confessions and course-wide chats (for example on Messenger). There’s a lot more public forums and discourse than I had when I was an undergrad.
I think there’s some good things and bad things about that. [On the one hand,] you can get support. [On] the other, if those public forums tend to be overly negative, that can be distressing for students. If they think that everyone’s feeling that way, the people that are reaching out for help who actually need help may feel that ‘oh, everyone’s like this.’ It’s okay to reach out for help, and I think sometimes that’s lost in some of those public discourse channels. I think communication is good in general in the community and in almost every area of life.
TT: Based on your website, it looks like you were a teaching assistant for many mechanical engineering courses. How did you end up teaching 6.006?
Ku: I’ve always enjoyed teaching because I always learn things when I teach and learn the material much better myself. I remember when I took AP Calculus in high school, I really enjoyed getting together with a group of my friends in the class and trying to explain things to get them to understand [the material] better.
As for the transition to Course 6, when I got to grad school, I noticed that for every problem I wanted to tackle, I needed to be able to process a lot of data or be able to run a simulation. It’s not that the interesting things involved computer science, but I had to be able to use a computer to do my work. I learned MATLAB as an undergraduate in MechE, and programming has always been interesting to me. But if you only get exposed to MATLAB, you don’t really know how to write a program to help you with your data without using proprietary software.
When I realized that I needed those skills, I took some Course 6 undergraduate courses as a graduate student. I took 6.046 as my first algorithms class. (Yeah, I would not recommend that to anyone.) I really liked the types of problems and the way of looking at the world in a discrete way that you could prove things about. You take a calculus class as your math [GIR] at MIT, but it’s not really a proof-based thing. Most actual math is about modeling the world and proving things about the world. I think courses in logic and proof-writing are useful, so I sought that out as a graduate student.
As I got into algorithms, I started doing research with [EECS professor] Erik Demaine, who does work in computational folding algorithms. I went to his lab for postdoc after finishing my PhD in mechanical engineering. That transitioned me to the computer science department where I applied for this teaching position and they gave it to me.
TT: In your time here, you have also been involved in computational geometry research and the advancement of the origami scene. What contributions to the MIT community are you most proud of?
Ku: An origami club at MIT, OrigaMIT, was around for quite a bit of time before I joined, but I became president of the club soon after I got here. There wasn’t such a big club presence throughout my undergrad, and it would just be me getting together with other friends in the community that folded every other month, maybe every three months.
I actually had some difficulty as a grad student as it’s a much different way of life. You’re working on research that is mostly self directed. I found that I had a lot more time, and I didn’t necessarily use that time wisely. I ended up putting a lot of extra time into making OrigaMIT more substantial. We moved the meetings from every other month to every week. I would give workshops and lectures on origami design, and I started a convention which has been going on for 10 years now. We usually have around 200 attendees. We opened it up to the general community, not just the MIT community. We started having enough student involvement and officer positions outside of me so that the club would perpetuate and live on its own, so I’ve been pretty hands off on the club for the past couple of years. I’m glad to have made something that can continue and I feel proud of that contribution to the MIT community.
In terms of research, you can look on my web page for the recent contributions. I don’t find [my research] making a huge impact on the MIT community, per se, but more of the community that’s interested in research.
The impact that I hope I’ve had the most on MIT is through my work on 6.006 and making it a better class for people and a little more standard. We’ve cut down on some material that probably shouldn’t have been there in the first place and expanded [the class] to teach material that I think is much more useful. We’ve also restructured the course to emphasize the right things so that students get to a middle-level understanding quicker. I hope that’s mostly what my legacy is.
TT: Could you expand more on the engineering applications of your research?
Ku: When I got to grad school, I went to a lab in mechanical engineering that dealt with making things at the nanoscale for optics-related devices and realized that in order to make things that are that small, you have to use lithographic processes, which is a very two-dimensional process.
These devices tend to be simple, so how can you get complexity in that kind of environment? One way to do that is to take these sheet materials you’re making and fold them up to make three-dimensional devices at that scale.
What is a fold? It’s as simple as two parts of the device that are physically moving relative to each other. Usually, there’s a constraint with how they can move relative to each other, and that’s a hinge or bearing that allows that movement. Anytime you have something that’s transforming, folding has an application there.
Where can you find an application for folding? I like to think that the single biggest revolution that we’ve had in my lifetime is the computer one, where you have a device reconfigurable in its state of electrons. We would never be able to put a computer in all the different possible states it could be, which means that there’s a huge design space of programs that you can write.
Building a general device that could be reprogrammed to do many things is what’s so powerful, so what if we could do the same with hardware? If I could download a software and a hardware update that reconfigures the matter in [my device] into a new one, I wouldn’t have to throw away anything. It’s kind of science-fiction-y and those things are far from being a reality, but there’s a lot of interesting research questions that go into how it’s possible to even reconfigure matter in that way.
There are physical problems and engineering problems we need to solve along with computational problems. A lot of the research Erik and I do deal with reconfiguration problems and how you can build these things with folding constraints, yes, but other constraints as well.
TT: With all that you’ve done, was there anything you didn’t get to do?
Ku: I do think that there’s a glut of students in Course 6 that probably shouldn’t be Course 6. I think most of the student body recognizes that using computers and knowing some computer science is an important skill to have in this day and age, but most kids at MIT should not be computer scientists; they just want to use computers to work on the projects that interest them.
We really should, at the Institute level, think of computer science more at the level of mathematics. It's [a tool] that can help with all of the other disciplines that you want. Most undergrads have to take calculus, differential equations, or linear algebra to support their major, but just as support. In the same way, taking classes in programming, discrete mathematics, and probability can be useful tools for doing any area of research or industry these days.
Much of the Course 6 curriculum right now is targeted toward Course 6 majors because we have so many of them. I really think it’s time for a lot of the undergraduate-level Course 6 classes to have multiple versions. I think [the current curriculum] shuts out a lot of students who would otherwise benefit from such kinds of courses: we don’t have, say, an algorithms class for non-computer scientists. I would have loved to start up such a course, but I didn’t have a chance to while I was here.
TT: How do you foresee the EECS department to evolve in the coming years, especially with its integration into the new College of Computing?
Ku: I honestly have no idea. If the College of Computing can help distribute computer science knowledge or training to non-computer science students, that would be a benefit. EECS is so inward focused on accommodating its own students that it doesn’t have the capacity to help the rest of the Institute.
Maybe this is an area where my opinion shouldn’t matter as much because I’m leaving, but I’m hoping that [the College of Computing] can do good things for allowing cross-disciplinary collaboration, hooking people up with the right [resources], and getting training for a broader range of folks. An influx of new faculty and funding should be good for those goals.
TT: What’s your latest origami project?
Ku: Oh geez. I haven’t designed any new origami things in a while, mostly because I’m enjoying my job more so it keeps my free time relatively occupied. Usually, the work I do is more about writing papers about folding and less about actual folding. I’m still involved in the origami community; I’m chairman for the US origami organization, OrigamiUSA, and I still stay involved with OrigaMIT.
Maybe after the term, when I transition to various two-week quarantines in various locations, I’ll start designing more. If any of your readers have suggestions on what to design, I’m happy to take them.
Original article appeared in The Tech on May 14, 2020.