Category: Pandas

  • How Search Works in Vector Databases โ€“ From โ€œTop Gunโ€ to โ€œMaverickโ€

    How Search Works in Vector Databases โ€“ From โ€œTop Gunโ€ to โ€œMaverickโ€

    โœ๏ธ ๐๐ฒ ๐€๐›๐ก๐ข๐ฌ๐ก๐ž๐ค ๐Š๐ฎ๐ฆ๐š๐ซ | #๐…๐ข๐ซ๐ฌ๐ญ๐‚๐ซ๐š๐ณ๐ฒ๐ƒ๐ž๐ฏ๐ž๐ฅ๐จ๐ฉ๐ž๐ซ ๐Ÿง  Why Do We Need Vector Databases? Traditional databases store structured data (rows & columns) and perform keyword or pattern matching.But AI systems donโ€™t think in keywords โ€” they think in meanings. For example: Thatโ€™s why modern AI systems like ChatGPT, RAG pipelines, and semantic search engines rely on…

  • Banned Pandas: Building Fast, Scalable Data Pipelines with DuckDB + SQL on Azure

    Banned Pandas: Building Fast, Scalable Data Pipelines with DuckDB + SQL on Azure

    โœ๏ธ By Abhishek Kumar | #FirstCrazyDeveloper ๐Ÿง  Why This Change Matters Our data pipelines were fast โ€” until they werenโ€™t.Large joins started crashing memory, inconsistent datetime types caused nightly job failures, and debugging hidden Python logic in Jupyter notebooks became a nightmare. So, we did something radical. We banned pandas.Every transformation now runs in DuckDB,…