Teakgyu: Hello, I’m Teakgyu Hong from the Agentic AI team, where I develop technologies required for Upstage’s products. 
I’m particularly interested in building AI-powered products that solve real problems for users.

Teakgyu: At Upstage, we focus on solving business problems using AI technology. Based on our mission, “Building intelligence for the future of work,” we develop AI products that go beyond simple automation to support complex decision-making and reduce operational costs, ultimately improving productivity for enterprises.

More specifically, I work on technologies behind Upstage Information Extract, a document information extraction solution that quickly and accurately retrieves key information from any document, and Upstage AI Space, an AI-powered document workflow automation platform that enables companies to perform various document-centric tasks more efficiently.

Teakgyu:  One of the most rewarding moments was when our product helped solve a problem for a large U.S. insurance company.

The client had previously attempted to address the issue using document processing solutions and GenAI products from global ICT 
companies, but none of them worked well. They eventually met us at a conference while searching for a solution.

Our team dedicated significant effort over a certain period to solve their problem using our product, and we were able to achieve a 
level of performance that satisfied the client. While the product still requires further refinement, it was incredibly rewarding to know 
that our AI technology had successfully solved a real user problem.

Teakgyu: Persistence.

With AI products, achieving 100% accuracy is often difficult. 
However, users still expect very high-quality results, which means we constantly go through many trials and failures.

Seeing colleagues persist through those challenges and ultimately achieve the goal has been one of the most valuable lessons for me.

Teakgyu:  I believe I’ve grown significantly in terms of productizing technology.

Before joining Upstage, I mostly focused on technical development. Problems and evaluation metrics were usually predefined, and my main goal was to improve scores.

In the real world, however, user problems are often ambiguous and vary from case to case, and clear evaluation metrics may not even exist. To turn technology into products that users truly need, I had to broaden my perspective, and that’s where I’ve grown the most.

Teakgyu: I want to build products that are used by people around the world and genuinely help automate their work.

Teakgyu: We typically hold a weekly online team meeting, and additional meetings are arranged as needed for specific projects.

We also try to document our work as much as possible and communicate through documentation so that discussions can be more 
structured and in-depth.

In addition, we sometimes hold offline workshops quarterly or semi-annually to strengthen trust among team members, review mid- to long-term goals, and improve how we collaborate.

Teakgyu: Previously, I worked at relatively large companies where I mainly focused on tasks assigned within my team.

Since Upstage is a startup, I’ve had opportunities to learn about what other teams are doing—such as the business, PR, and education teams—and to communicate directly with people across different functions. That was particularly impressive to me.

Teakgyu: During the hiring process, it’s helpful if candidates go beyond simply saying “I developed this technology” or 
“I used this tool.”

Instead, try to clearly explain what problem you were trying to solve, how you approached solving it, and how you evaluated the 
results. Sharing those experiences will make your work much more meaningful to interviewers.

Teakgyu: Building products with AI technology that are widely used by real users is extremely challenging.

However, it’s also incredibly rewarding because you can create real value with AI. If you want to be part of that journey, we’d love to 
see your application.