Jaeseong Choe

Data Engineer with a background in data analytics and computational communication research, building the pipelines and data infrastructure that turn raw discourse, behavioral, and business data into something analysts and researchers can trust.

As an engineer, I develop and maintain the pipelines that collect and load real-estate data at my current company, and build dashboards for analyzing it. Before moving into data engineering, I spent a year as a marketing data analyst. I consolidated a three-brand KPI dashboard onto Databricks, cutting maintenance effort by 90%, and built demand-forecasting models that reached 93% accuracy. I also designed a GEO (Generative Engine Optimization) framework for measuring brand visibility in LLM responses.

As a researcher, I apply NLP and topic modeling — STM, sBERT, and BERTopic — to media discourse and public opinion, with three studies published in KCI-listed journals and presented at national conferences, spanning psychiatric stigma in Korean newspapers to neo-nationalist discourse on YouTube. I also develop research software — PyBigKinds and RBigKinds, open-source Python and R packages that turn BigKinds news exports into analysis-ready data for text mining and topic modeling. Research remains a long-term direction, and I continue this work alongside engineering.

Across all of it, the throughline is the same: data infrastructure solid enough to trust, and analysis rigorous enough to publish.