byzer-lang

Why we included this project

Byzer, formerly known as MLSQL, is a SQL-like programming language for teams that want to run data pipelines and machine learning in one place rather than gluing together separate tools. Its design principle is that everything is a table: ingestion, transformation, and model steps all come through familiar SELECT-style statements, with the engine distributing the work over Spark. For data engineers in a Spark or Hadoop environment, that means shipping a data product without hand-writing Spark jobs. Byzer also includes a notebook and a VSCode extension for interactive development, and built-in algorithms plus an extension mechanism cover common modeling needs, so the same language can take a project from raw data to a trained model.