My internship at KRAFTON was a rare opportunity to focus entirely on research. I had the chance to collaborate with many talented colleagues, and the combination of freedom and access to computational resources made it an invaluable experience as an undergraduate student. As a first-time visitor to Korea, I felt immediately welcomed by the deep learning division, which made adjusting to a new environment much easier. Beyond work, company clubs such as climbing and orchestra quickly gave me a sense of community and allowed me to connect with people who shared my interests.

During my internship, I primarily worked on a project involving LLM agents for data synthesis. We developed an iterative agent that gradually evolved a dataset (or synthetic data generation pipeline) over time. This project taught me a great deal about designing and working with LLM agents. In addition, I contributed to a project on test-time scaling laws for LLMs, focusing on inference optimizations such as quantization and attention approximations, where I worked on quantization-aware RL training algorithms. These experiences not only strengthened my technical skills but also gave me valuable insight into conducting research in a structured environment.

Overall, my time at KRAFTON was both professionally and personally enriching. I am truly grateful for the teammates and friends I met, and I hope our paths cross again in the future.

The internship at KRAFTON AI was a precious opportunity where I could fully pursue the kind of research I had long dreamed of. From proposing new research topics to forming teams with fellow researchers and carrying out the entire research process, I was empowered to make independent decisions at every stage, which allowed me to deeply immerse myself in my work. The open culture that enabled free discussions across team boundaries greatly supported creative research endeavors. The team members were respectful of each other’s ideas while also providing sharp, constructive feedback that made a real impact on our paper projects.

Beyond academic projects, it was especially meaningful to participate in the Battlegrounds project. By being directly involved in the planning and development of AI products, I had the chance to apply my research to real-world services and to understand the challenges that need to be overcome to commercialize AI technologies.

All these experiences at KRAFTON AI have become invaluable assets in my career. I sincerely hope to cross paths with the team again in the future.

I first learned about the KRAFTON AI Fellowship through the MIT MISTI program and previous participants. That experience sparked my interest in KRAFTON as a company, and although I was born and raised in the United States, I’ve always wanted to start my professional career in Korea—so I decided to apply.

At the time of the exam, I hadn’t formally studied artificial intelligence, so I relied mostly on mathematical intuition to solve the problems. When I passed the third interview and received the final acceptance email, I was filled with excitement—but also a bit of anxiety and uncertainty.

The project I joined was related to LLM Agent research, similar to a recently published paper. I learned a lot from working alongside my team lead and fellow interns, who were incredibly knowledgeable and generous with their support. Getting hands-on experience with the latest LLM technologies, tools, and libraries was immensely valuable and will continue to benefit me even after returning to the U.S.

The culture within the department was very horizontal and autonomous. Even though I was the youngest member of the team, I never once felt like “just the intern,” and despite my Korean not being perfect, everyone treated me with equal respect.

The meals exceeded even my YouTube-fueled expectations. Some days, the quality rivaled that of fine dining restaurants, and the creativity of the daily rotating menu made lunch something I looked forward to every day.

Though brief, this experience gave me the confidence and courage to embrace challenges in any environment moving forward.