Jaewoo Song
Jaewoo Song


  • DailyLife

It’s been a while.

Starting last December, I have worked as a TA for a Natural Language Processing class in Boostcamp AI Tech which is provided by the collaboration of Naver Connect and Upstage AI.

I have been so busy conducting various tasks including my original works from my lab, so there has been no time to post.

And it is almost finished, allowing me to write about the personal record for this after a long time.

This TA activity was a very valuable change for me.

Although works couldn’t go smoothly and I got stressed sometimes, it was a meaningful time to meet/interact with various students and to review the basics and fundamentals.

The main three points here are as follows.

  1. There is not much difference between beginners and me than I thought.

    I was frustrated with the questions from novices which was quite sharp and deep many times. Of course, it’s been only 2 years since I started to study this field. I think I have been too conceited. When someone pointed out a mistake in my practice code and proposed a method I have never come up with(such as adding the pad tokens in front of the sentences to alleviate the long-term dependency problem when using RNNs…), I could not answer it immediately but also felt grateful. I realized that teaching is not only a process of giving one’s knowledge to others but also a process of growing together. I reflected on myself about my abilities and thoughts.

  2. I NEED to go to graduate school.

    Even though I have been working in an industry lab, it occurred to me that there is a huge gap between a person who is working or worked for an academic lab and me. In a school, we usually study and research more thoroughly than in a company which we just have to make an instant result. From this experience, I felt my passion for academic research increased and decided to go to graduate school as soon as possible to improve myself.

  3. No matter how a person is busy, he/she can handle many tasks at once. However, this is sometimes not the best option.

    For the past 3 months, I’ve done three things in parallel, including the company work, the personal research for a thesis, and TA stuffs. And I knew that I can conduct all of these things even if my time is limited. But I also noticed that doing as many things as possible does not produce the ideal outcomes. I just literally “DID” them and none of them was conducted satisfactorily. There has been not much progress in my company work, I failed my thesis since the test results were not desirable, and I have not been a perfect TA also. I concluded that I have to choose either quantitive benefits from participating in various works simultaneously only focusing on the number of things I do, not the quality or qualitative benefits by choosing and concentrating. And for me, the latter was the better choice, eventually.

In some ways, I only wrote a little bit negatively, but I think it was an invaluable experience and I don’t regret doing it.

If I had not tried it, I could not have got these lessons, after all.

Again I want to thank Upstage AI and Professor Jaegul Choo in the graduate school of AI, KAIST for trusting and hiring me.

I will continue to go on and make an effort more sincerely.