Summary
- Published 2025-08-12 · slug
boostcourse-sql · category "Computer science studies / Data, SQL, and systems" · source article recorded in Korean and English. - A review of SQL from an AI engineer's perspective, prompted by the realization that when Pandas groupby becomes slow, "that operation should finish in the database."
- Topics include JOIN and GROUP BY for training-dataset views; LAG, LEAD, and RANK window functions for time-series and ranking features in the database, such as time since the previous order; and subqueries and CTEs for breaking complex operations into stages.
- Workflow changes: moving feature engineering into the database, faster EDA with GROUP BY and COUNT, and communicating about problems in terms of schemas and partitions.
- Conclusion: SQL is a foundational skill for ML engineers.
Mentioned
- No organization or project links. The slug belongs to the Boostcourse series, but the body does not name a course; this needs confirmation.
Connections
- Writing
- Atomic note: Q: What are the two Coursera certificates, and are they listed? (Deep Learning start month and Machine Learning Specialization)
- Concepts: SQL · i18n and multilingual support · Running a technical blog · Data pipelines · Time series · Feature engineering
Local graph
Notes 1 · Links 0