Namhyeok: Hello, I’m Namhyuk Kim, a Data Manager in the Data team within the AI Product division at Upstage. 
My role involves creating the data required for the products being developed at Upstage. Recently, I’ve been focusing on how we can evaluate AI models and products more accurately.

Namhyeok: Currently, Upstage is developing two main products. The first is Information Extract, which extracts necessary information from documents into structured data, and the second is AI Space, which supports document-based workflows. 
The Data team builds various datasets required for developing these products, and we place particular emphasis on designing 
evaluation datasets that allow us to accurately measure AI product performance. Through this process, we also uncover key insights 
that help guide product improvements.

Namhyeok: One of the most memorable experiences was when I was assigned to a high-priority project for a new client. To meet both tight deadlines and high expectations, we formed a task force team and worked intensively day and night. In order to achieve the best performance, the training approaches and evaluation criteria often changed throughout the process. Each time that happened, I 
quickly reorganized the data and optimized the training and evaluation environments to support the team. When the client was 
satisfied and we successfully signed the contract, I felt a strong sense of accomplishment. Although it was challenging, it was also one of the most rewarding projects I’ve worked on, and it truly reflected the dynamic environment I was hoping to experience when I joined a startup like Upstage.

Namhyeok: I constantly learn from my colleagues through the way they approach their work. It’s not always easy for employees to 
maintain a strong sense of ownership and responsibility, but everyone here works with a high level of commitment, which motivates 
me greatly. There is no “free-riding” like you might see in group projects, and that creates an environment where everyone works with genuine passion. Thanks to this collaborative atmosphere, I also feel a stronger sense of responsibility and can fully immerse myself in my work. It truly feels like we are growing together through the positive influence of our teammates.

Namhyeok: The biggest area of growth for me has been expanding the way I view problems. Before joining the company, I focused 
mainly on solving the problem directly in front of me. Now, I try to understand the broader context and background surrounding the 
problem and think about more fundamental solutions. Rather than simply addressing immediate issues, I’ve learned to anticipate 
potential challenges that may arise in the future and prepare for them. This shift in perspective has helped me propose solutions that 
create greater long-term value.

Namhyeok: During my interview at Upstage four years ago, I was asked what my goal would be ten years into the future. At the time, 
I answered that I hoped to become one of the leading AI data experts in Korea within ten years. That goal still holds true today, and I 
have about six years left to achieve it. While others may think differently, when I reflect on my current pace of growth and the 
experiences I’ve gained so far, I believe it is a goal that is absolutely achievable.

Namhyeok: In AI development, data must always come first, so it is important to anticipate needs and stay one step ahead. If training data is not prepared in time, models cannot be trained, and if evaluation data is delayed, models cannot be properly validated. For that reason, it is essential to understand the overall workflow, adjust priorities depending on the situation, and deliver the best possible 
results within limited time. Even after datasets are released, the work does not end. We continue to communicate closely with 
engineers who use the data and continuously improve data quality and evaluation methodologies.

Namhyeok: When I first joined Upstage as my first job, I had many worries as a new employee. I often wondered whether I could 
perform well without causing trouble for my colleagues, and how I could prove my value. I constantly questioned myself—“Is this 
approach okay? Am I doing this the right way?” However, after experiencing several successful projects with the support of my 
teammates, those initial anxieties quickly disappeared. I gradually gained confidence and was able to work at my own pace. 
Looking back, everyone feels nervous and uncertain at the beginning, but eventually you learn by facing challenges directly. 
In many ways, company work is not so different from the experiences you’ve already had. Even now, new challenges continue to arise, but unlike before, I approach them with confidence and a much more relaxed mindset.

Namhyeok: Above all, we value the ability to solve problems with a data-centric mindset—an approach often referred to as 
Data-Centric AI. Therefore, the most effective way to demonstrate your strengths during the resume, portfolio, and interview process 
is to show how you have applied this mindset in practice. Even with simple modeling experience, you can highlight how you defined 
and solved problems from a data-centric perspective. For example, creating synthetic data to address insufficient training data, or 
designing evaluation criteria and building evaluation datasets to validate models when such data did not exist, are both strong 
examples.

Namhyeok: Many people who study AI naturally set their career goal as becoming an AI/ML engineer or researcher. 
I thought the same at first. However, after working in the AI industry, I realized that there are many other roles and opportunities 
available. I hope applicants will explore different paths with a broader perspective and discover the career that suits them best. The AI field offers far more diverse opportunities than many people expect, and there are many places where you can apply and grow your 
capabilities.