Building a Fast, Memory-Efficient Hash Table in Java (by borrowing the best ideas)
One day, I ran into SwissTable—the kind of design that makes you squint, grin, and immediately regret every naive linear-probing table you’ve ever shi...
Software Engineering for Data Scientists, Part 1: Pydantic Is All You Need for Poor Performance Spaghetti Code
leehanchung.github.ioSaved July 5, 2025
Summary
Learn why using Pydantic beyond API boundaries leads to 6.5x slower performance and 2.5x more memory usage. This post reveals the serialization/deserialization debt anti-pattern that creates performance bottlenecks in Python applications and shows benchmarks comparing Pydantic with native dataclasses. Discover when to use Pydantic (data validation at service boundaries) and when to avoid it (everywhere else).
