I am a PhD Candidate at the Leiden Centre for Digital Humanities and the Leiden Institute for Area Studies at Leiden University, specializing in Korean Studies.
My research applies computational tools to Korean history, with a focus on computer vision, text mining, and large language models. I aim to enhance our understanding of Korean history — from the colonial period through the modern era — through innovative digital methods.
Positions
Current
- PhD Candidate — Digital Humanities & Korean Studies
- Researcher — Leiden HumAN
Previous
- Junior Research Fellow — Kyujanggak Institute for Korean Studies
Education
- M.A. Asian Studies (Korea Focus)
- B.A. Korean Studies
Dissertation Research
My dissertation explores the application of computer vision, text mining, and large language models in understanding Colonial Korean history through two distinct case studies:
Typography & Printshops in Colonial Korea
I examine how Korean printshops during the colonial period developed recognizable visual "signatures" through type design and printing practice. Instead of treating typography as abstract art, I treat it as material culture shaped by technology, labor, and politics.
My core work uses large-scale image analysis and deep learning to identify recurring visual traits in printed pages: stroke weight, modulation, alignment, margins, layout, density, and compositional rhythm. Where earlier scholarship relied on close reading of a few exemplars, I map patterns across tens of thousands of pages.
These signatures reveal workshop practice, standardization processes, and the industrialization of Korean textual culture under colonial rule.
Advertising & Consumer Culture in Colonial Korea
I study the rise of mass advertising in colonial-era newspapers as a visual and social system. Using computer vision, I first segment and extract hundreds of thousands of advertisements from page images. But the real methodological leap comes next: I use LLMs to classify and tag those ads with rich metadata at scale.
The contribution isn't just scale. It's the move from scattered examples to a structured dataset of visual culture. Advertising becomes quantifiable as historical evidence, while still grounded in close reading and critique.