TA'ing 2.009 and 2.744
TA’ing 2.009 and 2.744 has roughly been my full-time job for the past year or so, and 2.744 just wrapped up. This has been an incredibly rewarding, frustrating thing to do, and I’ve learned a lot. I’ll try to expand on that, but it’s hard to enumerate all the things that I’ve learned, because so much of it is soft skills.
First, some background. I was a double-major as an undergrad, 2A and 6-3, so I took 2.009 in Fall 2016, graduated Spring 2017, TA’d 2.009 Fall 2017, TA’d 2.744 Spring 2018 (I haven’t ever taken 2.744). Both classes are taught by Professor David Wallace
2.009 is MIT’s MechE capstone class, “Product Engineering Processes”. Students are assigned to teams of about 20, initially split into two halves. The task of the semester is to design a product for the year’s theme, and present an alpha prototype at the final presentations. The year I took the class, the theme was “Rough Tough & Messy”, and when I TA’d, the theme was “Super”. Themes are generally intended to be fairly open-ended, but there have been some slightly more focused ones in the past, like “Food”, from 2010. 2.009 is very structured, with milestones roughly every two weeks, and most of the lectures focused in engineering process - ideation, brainstorming, prototyping, testing, consensus-forming. I’ll talk more about my experience taking 2.009 in a future blog post though - this one is about TA’ing 2.009.
2.009 overall is characterized by its insanely high production value. Skim through the final presentations from 2017 to get an idea of that. There are usually 4 TAs for 2.009, and we generally handle all of the media associated with the class - nearly 100 thousand photos and many terabytes of video footage, not to mention dozens, if not hundreds, of graphics. It’s a monumental undertaking - I probably worked at least 60 hours a week on average, more than twice as much as I spent on 2.009 as a student. As I mentioned, it was incredibly rewarding, frustrating, and a distinct learning experience.
2.744 is a graduate product design class, taught every other year by Professor Wallace. Students form teams of about 6, and the primary task of the semester is to design a product for the semester’s client - this year, the client was 5 Wits, a production company by MIT alum Matt Duplessie, and students were tasked with designing gags/games to be used in a new concept that 5 Wits has been exploring. 2.744 is skill-focused, with many lectures on sketching, rendering, prototyping, and general design lectures. There are fewer milestones throughout the semester than 2.009, and teams are much more independent.
2.744 is a much lower-key class than 2.009, but there are still a number of high production value components, and the experience as a whole is still very tightly designed. Again, TAs are generally responsible for the media associated with the class in addition to general class prep. However, the scale is very different - only about 10k photos throughout the semester, as opposed to the roughly 100k from 2.009. The only video components are documentation, there’s no creative video editing to be done, and the number of custom graphics is also smaller.
Learning - hard skills
I didn’t learn many new hard skills through TA’ing 2.009 or 2.744, but I definitely sharpened some existing skills. In no particular order:
- I got a ton faster at photo editing. With the sheer volume of photos to edit, I got faster out of necessity. That involved memorizing and committing to muscle memory most of Lightroom’s keyboard shortcuts. Check out the 2.009 and 2.744 galleries for some examples.
- I also developed a stronger sense for organizing photos. Most of my photo sets previously have just been ordered chronologically, but photo sets for 2.009 and 2.744 were rearranged creatively for visual interest, telling a story, or visual puns.
- I got even stricter with my folder and naming conventions, again, out of necessity. Unfortunately for 2.009, this effort was confounded by some of the other TAs not always being as strict…
- Memory card organization. I numbered all of our dozens of memory cards, and tried implementing a system for “to dump” cards where the contents were written on a post-it so that the photos could be appropriately sorted and labeled. That too was confounded somewhat by other TAs being less strict…
- Physical camera organization and operation - in keeping with the theme of organization of large quantities of media/things, I took point on our camera organization, charging setup, and prepping cameras for class, which involved selecting lenses and ensuring each camera had a fresh battery and empty memory card. It doesn’t sound like much, but when you do it 3 or more times per week for 4 months, it becomes notable.
- For 2.009, I didn’t actually take or edit that many photos, relatively speaking - I usually shot video during lectures, milestones, and events, and organized all of that media.
- Video editing itself was something I had done only minimally before 2.009. During 2.009, I was the primary editor for two videos: Teardown and the quadcopter video. Both of those probably deserve their own post, but basically, video editing is hard. I always felt like I needed more footage and more time. Getting high-quality footage is really difficult in the context of 2.009, where most things are filmed run-and-gun style, without enough time to think through things. 2.009 also uses a bunch of different cameras, and it’s very difficult to make them all look nice when edited together.
- Flying a quadcopter was something I’d never done before 2.009, and though I’ve crashed (twice), I consider myself fairly decent at it now. It really gives a unique perspective and raises the production value.
Celebrate MIT: Quadcopter 2017. Probably my proudest artistic achievement.
- Bash scripts are something I’ve used before, but not as extensively as I used them in 2.009/2.744. In 2.009, we had a video editing script that Victor wrote a while ago, which you could give a video file, in and out timestamps, and a new name, and it would automatically clip that section from the videos, with a fade in and fade out, and rename it. This was extremely useful for milestones, where I would have to produce 48 video files (6 ideas per team x 8 teams) in just a few hours. For 2.744, I had to systematically rename tons of photos - for in-class assignments, I either scanned paper submissions or took photos of every student’s submission. I had to edit each of these and then rename them to the student’s kerberos, which was easily scriptable.
- To help with the bash scripting, I utilized Google Sheets pretty heavily - columns for different fields and a final column to merge them together appropriately to the command to be invoked. Obviously, bash itself or some other scripting language would be capable of this, but it was easier to visualize and edit in Google Sheets.
- To incrementally back up the many, many terabytes of photos and videos, I used rsync, which is a cool tool that I haven’t used before. It definitely beats individually figuring out which folders or files are on which drives.
- Spreadsheets as an organizational tool. I’ve used spreadsheets before, obviously, but usually in the normal contexts spreadsheets are used - numerical data. I’ve also used them for puzzle-solving, but that’s a different story. Using them for organization is something that makes a lot of sense to me - organizing the camera setup for build challenge was the prime example of this, where I coordinated 42 different cameras to record or photograph simultaneously, and organized all of that data afterwards. It took me nearly 3 days just to dump and back up all of the files from build challenge.
- I’ve always had a bit of an eye for web design, but it was pretty unrefined. This (warning: language) was generally my idea of a nice-looking website (and, as you can tell by this website, that hasn’t changed much). What has changed is that I’ve refined that vision a bit, and improved my familiarity with CSS & JS by working on the 2.009 and 2.744 websites (though they still need a pile of work, if not a redesign from scratch).
- That’s all well and good for web dev, but I’ve also gotten exponentially better at Illustrator, moving from nearly completely incompetent to only somewhat incompetent! Still have a lot of work to do here, but I’m at least more likely to open up Illustrator to accomplish something than most other programs nowadays.
- Finally, designing a lecture, or a curriculum, is another interesting challenge that I hadn’t really considered prior to TA’ing. The goal is to convey information in an organized, memorable fashion, and there’s a lot of things that go into that. Organizing topics in a lecture, balancing the amount of information conveyed vs the length of the lecture, and using relevant examples or exercises all tie in to conveying information. I had the opportunity to give a lecture and a half on web design for 2.744. Luckily, I had Victor’s slides from 2016 as a basis for my lectures, but was able to update them, reorganize them, and put my own spin on them. I got a lot of positive feedback from the students, and was overall very happy with the lectures (aside from some verbal tics that I’d like to work on for future public speaking opportunities).
- For the physical aspect of making things, I mostly served as a consultant for teams in both 2.009 and 2.744, but I did get a chance to flex my own skills a bit. In 2.744, I was in charge of 3D printing 219 soffits for the teams to use in their scale models (I know the exact number, of course, because I used a spreadsheet to organize all of this). That involved quite a bit of fiddling with the printers and print settings to achieve a good balance of print speed and reliability - at the end, I got the speed up to about 20 mins/part, down from about an hour per part prior to optimization, with about 95% reliability.
- At one point in the 3D printing process, casting crossed my mind as a potential way to increase the production speed, so I tried casting a few parts. It worked alright, but I didn’t spend enough time dialing in my process to make perfect parts. Still, it was my first experience casting anything, so that was cool.
Learning - soft skills
How to make Nice Things™
- This involves a huge variety of skills, which I’ll go over in future sections, but 2.009 especially has a lot of Nice Things™, from media to lectures to activities. One of the biggest endeavors every year is build challenge, where we literally built a mountain in the middle of Kresge.
Designing an experience
- As I mentioned at the start, 2.009 can be characterized by an overall insanely high production value. This was certainly on display for the videos, and something I’d had less experience with beforehand. All of the videos went through multiple review stages, and the editing was always tight, the content high-quality. Putting together a piece of work with such a high effort-to-length ratio is something I had very little experience with before 2.009.
- One of the aspects of a high production value is that the experience is excellent - everything is well-organized, and things are delivered or published in a timely fashion. This was definitely true for photos, which were usually published to the course website within 48 hours. For 2.744, with the lower photo volume and less things to do, I usually published photos within 8 hours. I actually have fluctuated on my posting speed - in freshman & sophomore year, I would post photos of big events within 12 hours, usually, going so far as to pull all-nighters editing, just so I could have photos out by the following morning. Through college, my editing preferences tended towards higher-quality, lower-speed editing techniques. 2.009 and 2.744 refined that further to a balance of the two, where I could maintain nearly my highest standard of quality, but still get photos out quickly. Unfortunately, photo editing seems to be an \(O(n^2)\) type of activity, so big events still take me a while to edit.
- As an extension of the high-quality experience that 2.009 and 2.744 had, lecture preparation was a surprisingly impactful thing to me. In 2.009 and 2.744, there are frequently very involved lecture demos, handouts, mini-quizzes, etc. The TAs would set up the lecture hall beforehand for the students before they came in, laying out the handouts at each seat, setting up the demos. For the handouts, we would not only set handouts at every seat, but would organize them precisely so that each student had a “place setting”. This, I think, was an underappreciated but crucial part of the 2.009 and 2.744 experience. I actually walked into a math class that I was thinking about taking, and when we had to pick up handouts from the front of the class, it felt distinctly lower-quality. By ensuring that the student experience was so polished, I think 2.009 and 2.744 helped ensure that the student projects reached a similar level of polish.
Another aspect of the production value of 2.009 and 2.744 was the consistent theming and design language of everything. Most classes actually achieve this, but only by virtue of having no theming and the word editor default fonts for everything. This is something that I actually felt pretty strong in even before TA’ing - my personal brand/website has always been the same grey/blue color scheme, and my logo has always been the block “LC” of the favicon.
- 2.009 had the theme of “Super”, and a design language consisting of randomly-arranged triangles in the team colors, evolving to an 8-bit graphic style and a retro-gaming theme for the final presentations.
- 2.744 had no “theme” per se this year, but the graphic language consisted of black background, outlined “2.744”, and purple and orange highlights. All of our website and graphic theming followed this, and frequently was angled consistently to add visual interest. Overall, a bit less complex, because there were fewer things to theme, but a consistent design language helped drill into the students the importance of design.
Working on a team
- One of the most difficult aspects of taking and TA’ing 2.009 was working on a team. I consider myself a decent team leader, and identified some points to work on during 2.009 and while TA’ing. One of my strong points as a leader is that I have a wide pool of knowledge and experience to draw from, and thus can be efficient and decisive. The flip side is that I’m not very good at giving other opinions a fair shot, which can lead to friction. That said, if there’s a matter that I’m indifferent to, or not knowledgeable in, I can be very hands-off.
- An aspect of this flaw of mine is that when others on the team are working on things I’m not directly involved in, I probably don’t give them the credit they deserve, because I don’t clearly see all of the work to get to the end result, only the result.
- One other aspect of this is the value of specialization in tasks vs balancing workload between team members. Though it was never really voiced, I felt like an underlying tension in the 2.009 TA team was workload balance. At times, any of the 4 TAs might be working much harder on a particular task than others, or working on an undesirable task, and anytime I had to stay later than everyone else, I generally felt annoyed. One way to solve this would have been to have everyone work on every task equally, but that’s not feasible - things get more efficient when you specialize. This can also lead to a general lack of appreciation of everyone else’s work. I under-appreciated the other TAs for their work on miniquizzes, but I also felt like my work on camera organization went under-appreciated when the other TAs didn’t follow the system I tried to establish.
- Compounding all of these issues was, of course, sleep deprivation. I get more irritable when I’m extremely tired, and I don’t think there’s any way around that except to acknowledge it and try to avoid scenarios like that. Of course, that’s impossible in a setting like 2.009 when there’s so much work to do, so I’m sorry to my fellow TAs for all of the times that it was late, I was tired, and I was irritable.
Working with a manager
- The other aspect of working on a team is working with a team leader, or manager. I've had surprisingly few times when I actually worked for someone on day-to-day tasks. In all of my internships, tasks have generally been assigned in much larger chunks, and the process of breaking those tasks down into hourly or daily activities was always up to me, as well as the scheduling of those smaller pieces. Working on 2.009 and 2.744 was completely different - tasks were assigned in small pieces (usually less than 2 days of work), but were generally assigned much more last-minute. In any given day or week, I rarely knew what I would be doing more than a day in advance, and this made scheduling and load-balancing very difficult. I had no idea how much that would affect me, but this was absolutely my chief complaint with 2.009. This was significantly improved in 2.744, mostly because Audrey knew the class very well, and was very proactive in asking for specifics on upcoming lecture and tasks.
- This was also my first time working with a manager that wasn't working on the same sorts of things as I was - in all of my internships, I was working in the same codebase as my mentors, so they had a really good idea of what I was doing, how difficult it might be, and how long it took. Working with David was pretty different from that - David was focused on lectures, and almost never was working on the same things as the TAs. This led to more pain and compounded the lack of control over scheduling. It was frustratingly frequently that David would offhandedly ask for an additional task to be done that would take at least a few hours as if it were nothing. The most painful example from 2.009 was when I was working on build challenge footage.
The lack of acknowledgment of my opinion on how long it would take to accomplish the task was incredibly frustrating, and wasted both my sanity and the editor's time.
I was working on organizing all of the footage, sorting through all 60 hours of footage to extract the 1% that was useful. This was a pretty monumental task, and I estimated it would take me about 4 solid days of work to get it done (after 3 days of work just transferring, backing up, and organizing the footage). Our video editor asked on a Thursday for an estimate of when he could get the processed footage, emphasizing that he needed it ASAP to get started on editing. I said I could get it to him by Tuesday for sure, or probably even Monday afternoon. He said he absolutely needed it by Monday to get started in time, and I responded saying that I'd try my best to get it to him by 9am Monday. David replied, not acknowledging my email, and promised it would get done by 5pm on Sunday. I worked straight through the weekend, sleeping only a few hours Friday and Saturday night, and sure enough, come 5pm Sunday, I didn't have it done. I literally couldn't get it done by the deadline David had set, and as a result, the video editor had to make an extra round trip to MIT to collect the footage (I nearly made it, finishing around 9pm on Sunday).
- At times, David would have feedback for us, and almost always, this feedback was given to the TA team as a whole. This was fine when the feedback was universal, but frequently it was more personalized. At one point, he critiqued a photo gallery that I'd put together, noting that it didn't tell a cohesive story, unequally featured certain members of the class, and a few other points. This was all fair critique, and I got better at editing galleries as a result (those weren't even things I was thinking about beforehand), but having the feedback given so publicly wasn't something I appreciated.
- This list has been mostly negative so far, but David was far from a bad manager - he was an excellent role model and absolutely spearheaded the high quality that I appreciated so much from 2.009 and 2.744. He was always available to give advice on how to do things. It's a bit hard to explain, but just by being in the presence of such a hard worker, with such high expectations and a high level of polish completely raised the bar on everything I'll do in the future. David was so incredibly willing to put in 110%, and put himself out there, that the TAs couldn't help but follow.
TA’ing 2.009 and 2.744 was, as I said at the start, incredibly rewarding and, at times, incredibly frustrating. I am eminently grateful for the opportunity to work on such a large-scale, high quality activity, and my sense of experience design and visual design has grown exponentially thanks to the influence of such a high-expectations, talented group of people. The quadcopter video I made for 2.009 is one of my proudest artistic achievements, and the web design lectures I gave for 2.744 were two of my proudest teaching/public speaking moments. Through all the stress and points of improvement I’ve highlighted, there has been also an extreme trust and level of respect for the TAs, for which I’m grateful. Thank you to David and my fellow TAs (Grace, Chuck, Georgia, Audrey, and Josh, plus bonus TAs Dabin, Victor, and Geoff) for an incredible year.
If any future TAs or students happen across this post, feel free to ping me through the usual channels for any extra advice or opinions. Thanks for reading :)