Course overview and setup

On this page

This module introduces the course and sets the expectations for how a Data Steward should learn the product.

Learning objective

Learn the steward role focus: observe the instance, investigate a problem population, make a stewardship decision, and verify the result.

What this role owns

Based on the documentation, the Data Steward role is naturally aligned to quality-focused, operational product work such as:

  • finding suspicious records;
  • isolating bad data patterns;
  • reviewing validation failures;
  • reviewing duplicate candidates;
  • confirming that the visible issue is gone after a fix.

Built-in AI agents

Before you begin

  • Use a dev or test environment first.
  • Pick a business domain with enough data to make search, filters, validations, and deduplication meaningful.
  • Make sure learners know they are not expected to create architecture decisions in this course.
  • Workshop format: 6 modules, 60–90 minutes each.
  • Self-guided format: 2 weeks with one module every 1–2 days.
  • Shadowing format: pair the learner with an experienced steward for modules 4–6.

Success criteria

The learner can describe the steward workflow in one sentence: search → filter → validate → deduplicate → verify.