My first two weeks were mostly spent getting up to speed on programming. Which isn't something that comes naturally to me, since I'm not a CS major. But I'd chosen the machine-learning side of the project on purpose. I thought it was the most intriguing part and most applicable to any situation or study. I wasn’t expecting it to be easy. I just wasn’t expecting a 25-hour video on how to use PyTorch.
At first, even that was fine. I was motivated. I'd flown across the world to learn new things, with state-of-the-art materials right in front of me. But around hour 8 of the video, I found myself not really paying attention anymore but instead just watching it to look busy in front of everyone else.
What I didn't understand at the time is how quickly that state feeds on itself. With no clear next step and no real momentum, the days started to blur, everything felt monotonous. The more time went by with nothing to show, the more I avoided the people who might ask how it was going. And the less I talked to them, the less direction I got, which meant even less to show. I was isolating myself, and every day of avoiding a difficult conversation made the next day harder to have it. Looking back, I don't think I was lazy or out of place. I think I'd fallen into something a lot of people fall into and I just happened to be conscious of it happening to me.
The one consistent break in the routine was my labmate Hieu, a master’s student at the university. Every so often he'd bring me along when the team went to watch the fiber being extended. The physical, hands-on side of the project that’s actually in a lab that I'd never have seen from behind my screen. He didn't have to. But those trips were the only times the work felt bigger than my own small lane, and they were a glimpse of how much was actually happening around me while I sat boxed in mine.
What stuck with me wasn't only what I saw on those trips. It was that Hieu took the initiative at all. He didn't have to fold me in. He just did it. And slowly that shifted something. Seeing how much was happening around me, and being pulled into it by someone who had no reason to bother, made me want to actually be part of it instead of watching from the edge. It was the main reason I had the confidence to speak up and know specifically how more involved and in what I wanted to be rather than blend into the desks.
Before it got too out of hand, I worked up the courage to do exactly that. Which, as easy as it looks on paper, really wasn't in the moment. Another version of this summer could easily have been me stuck in that same loop, never getting out. But as soon as I asked for more work and guidance, it was given to me.
I went to Prof. Tung, who pointed me to Dr. Trung. Dr. Trung actually sat down and looked at what I'd been working on, then planned around it; moving me onto data collection, which meant I was now in the lab itself instead of the office. From there, things built much more quickly. Because of those trips with Hieu, I understood the whole pipeline and not just my slice of it, and spending real time in the lab meant I understood the manufacturing side as well as the software, rather than knowing one and being lost on the other. That turned out to matter more than I expected. There were real problems holding up the project, and because I could see the whole process end to end, I was often the one who could pinpoint where it was actually breaking and what to change. A lot of the fixes were small tweaks, but they added up, and being able to make them is what made me useful. Eventually it put me in a position to lead the work: a small team of two, plus Prof. Tung. I'm now the first author on the paper currently in development.
Looking back, the hardest part was never asking for the help itself. People turned out to be remarkably willing to bring me along, make room for me, and point me in the right direction, but none of it could happen while I stayed quiet. The hard part wasn’t finding people willing to help. It was letting them see that I needed it. Whatever I got out of this summer, the paper included, happened because people were generous enough to bring in someone who had zero experience and who would be gone in a little over eight weeks. I’m very grateful for it.
I went to Prof. Tung, who pointed me to Dr. Trung. He sat down, looked at what I'd been doing, and moved me onto data collection, which meant I was finally in the lab instead of the office.
And being inside the lab was very different from the office. Instead of filling time and passively learning, I had real things that needed to be done on a timely basis: I'd record data, see what wasn't working, and dig into why the model was off. A lot of times the problem traced back to the fiber itself; so I'd flag it to Hieu, he'd fix the fiber, I'd improve the recording setup, and the next batch of data would come back cleaner. Each pass made the next one better. I now had a reason and necessity to get to the lab as early as I could and spend time there. Talking to anyone that went by instead of avoiding them.
That's also when spending time on both sides paid off. Everyone knew their own part better than I did, but I'd picked up a rough feel for the whole pipeline, and that turned out to be extremely helpful for being able to keep the pieces connected. That's how I ended up coordinating the work, a small team of two, plus Prof. Tung.
We're writing the paper now, aiming for an IEEE submission. At its core, it's an advance on how Prof. Tung measured touch on these fibers back when he first worked on them. I'm the first author. And I’m very grateful for all of the amazing people I met at VinUniversity and the opportunities they gave me.