Summary
Dongkyu Choi is a Seoul-based machine learning engineer with 9 years of experience building production-grade ML systems across fintech and platform companies. He designed Modelier, a feature meta management library that replaced tens of thousands of lines of SQL and helped consolidate legacy metadata into a single source of truth, and built scalable PySpark pipelines and custom deserialization to process high-volume feature data memory-efficiently. His models have driven measurable business impact—boosting home loan conversions to generate an estimated $1.3M weekly and doubling conversion rates for other key products—while also shipping targeting, income-prediction, and DSP/CTR solutions. With a background in economics and financial engineering from NYU and prior work at the Bank of Korea, he combines rigorous quantitative modeling (ASR, LSTM nowcasting, FX prediction) with hands-on engineering and product sensibility. He’s an all-time learner who routinely turns complex data and legacy stacks into reusable, developer-friendly tooling that accelerates modelers’ workflows.
9 years of coding experience