Data pipeline

A sequence of processes that moves data from its source to a usable output.

A training pipeline may ingest recordings, synchronize streams, select clips, attach annotations and assemble releases. Each transition can introduce errors. Review requirements should specify what is checked, which version is checked and what happens when an item fails or needs further judgment.

In practice

An automated file check catches a missing camera stream; a rubric based review checks whether a task description matches the recording.