The Korean language is used by more than 80M people and is increasingly gaining popularity due to K-dramas, K-pop, variety shows, and games. However, despite its popularity, it is a minor lanuage in natural lanyage processing research as most research is carried out in English.
During her graduate studies at Seoul National University, Lucy learned that many open source packages for the Korean language were often difficult to access. So in 2014, she developed and released KoNLPy, an open-source Python package for Korean natural language processing, focusing on easy access, usability, and scalability. The package garnered many users in Korea and is still widely used for production and educational purposes. She furthured her passion for the Korean language while working on Papago, South Korea's tech giant Naver's machine translation application. She built an honorific-sensitive machine translation application of Naver in 2017, which gained popularity for many people learning the Korean language. In 2020 Lucy co-founded Upstage, where she keeps developing AI models to make AI beneficial to many people.
Upstage is an AI solution company working on computer vision, recommender systems, and natural language processing. One of the efforts they recently made for Korean language processing is to lead the initiative for KLUE, a Korean language understanding evaluation benchmark. This is a big deal since the released dataset is one of the first and most extensive Korean datasets free from copyright license issues. Lucy participated in creating data for the machine reading comprehension task. Upstage has earned 8.8 billion won in revenue in just eight months since its founding by providing AI solutions. They will be releasing reusable AI packs in the near future.