Datanatomy
VideosBlogCodeAboutLinks
/

    GitHub
    DatanatomyCode is MIT on GitHub
    YouTubeTikTokInstagramThreadsGitHubRSS
    Datanatomy
    VideosBlogCodeAboutLinks
    /

      GitHub
      36 videos · page 2 of 3

      Data engineering, taken apart.

      Long explainers, 30-second cuts and quick tips. Open one for the player, the key moments, the full transcript and the code.

      All 59Machine learning 14AI engineering 4Data engineering 35Methodologies 4Software engineering 2

      13–24 of 35 vertical videos, page 2

      Series
      Length
      Watch on
      Sort
      • Frame from Batch vs streaming: 2 seconds vs 12 hours: Batch vs stream, the same fraud rule, nightly or live0:39Data engineeringBatch vs streaming: 2 seconds vs 12 hours
      • Frame from RAG ingestion: re-embed only what changed: RAG ingestion, re-embed only what changed0:35Data engineeringRAG ingestion: re-embed only what changed
      • Frame from Lake vs warehouse vs lakehouse in 30 seconds: Where data lives, lake vs warehouse vs lakehouse, in code0:37Data engineeringLake vs warehouse vs lakehouse in 30 seconds
      • Frame from Three data jobs, one dataset: Three data roles, data engineer vs data scientist vs ML engineer0:38Data engineeringThree data jobs, one dataset
      • Frame from 5,500 tiny files, 20x slower: Small files, why thousands of tiny files slow queries0:34Data engineering5,500 tiny files, 20x slower
      • Frame from Data contracts stop silent $0 days: Data contracts, catch upstream drift before load0:38Data engineeringData contracts stop silent $0 days
      • Frame from Why data platforms use three layers: Medallion layers, bronze, silver, gold0:36Data engineeringWhy data platforms use three layers
      • Frame from Stop reloading a million rows: Incremental loads, watermark, lookback, MERGE0:35Data engineeringStop reloading a million rows
      • Frame from SCD Type 2: stop rewriting history: SCD Type 2, keep dimension history with DuckDB0:37Data engineeringSCD Type 2: stop rewriting history
      • Frame from CDC catches every change: The change log, change data capture with sqlite triggers0:39Data engineeringCDC catches every change
      • Frame from Undo a bad write with Iceberg: Apache Iceberg, undo a bad write with one rollback0:34Data engineeringUndo a bad write with Iceberg
      • Frame from One rerun, twice the revenue: Idempotent loads, make reruns safe with DuckDB0:34Data engineeringOne rerun, twice the revenue
      ← Previous
      1. 1
      2. 2
      3. 3
      Next →
      DatanatomyCode is MIT on GitHub
      YouTubeTikTokInstagramThreadsGitHubRSS