Professional learning path

Data Quality Engineer

Define and review quality evidence for a dataset or model input.

Prepare

Design a recoverable reconciliation run

Run the reconciliation kit, then repeat it with the same inputs. Define how the production workflow should handle retries, late settlements, monitoring, and a failed downstream handoff.

Start with the browser demo or download the local kit. For the workspace exercise, you need approved access to the relevant Genedata capabilities and a reviewer for your output.

Practice tasks

0 / 4 completed

1
2
3
4
Knowledge checkpoint

A run fails after creating its review queue. What should a retry preserve?

Notes and progress are saved on this device.

Apply in your workspace

Catch problems before the handoff.

Define and review quality evidence for a dataset or model input.

1

Agree the critical fields, expected distributions, and downstream acceptance criteria with the owner.

2

Create repeatable validation checks and record the baseline or evaluation set used.

3

Run the checks before publication or promotion and investigate material differences with engineering.

4

Share the result, exceptions, and accountable owner; refresh the baseline only after reviewing the reason for change.

Expected handoff

A quality assessment with checks, baseline references, exceptions, and ownership.

Explore your platform capabilities

Measure your progress

  • Quality issues found before publication
  • Time to resolve an issue with the responsible owner

Use AI to propose edge cases and explain failed checks; have domain owners approve what counts as acceptable data.