Qualification engineering against independent oracles
Author independent scalar/grouped expectations. Design negative cases and exact scope.
01 The syllabus
Inside the course.
7 lessons / read + practicePublic lesson summaries. Join the design-partner program to discuss early access. Enrolment and assessment policy remain pending approval.
- 0160 min
Establish oracle independence
Separate expectation authorship and capture from the implementation under test.
- 0270 min
Freeze a useful catalogue
Define check identities, coverage, tolerances and business assumptions before acceptance.
- 0370 min
Handle precision deliberately
Contrast exact SQL decimals, semantic projections and displayed rounding.
- 0475 min
Test groups and context
Cover complete key sets, filter states, nulls and empty results rather than one grand total.
- 0575 min
Plant realistic defects
Detect a dropped group, duplicated batch, changed grain and wrong displayed value.
- 0670 min
Investigate without moving the goalposts
Use receipt supersession, strict evidence and explicit correction/refreeze policy.
- 0760 min
Defend the claim
Present scope, results and limits to an independent assessor.
Your learning evidence
reproduce a baseline and correctly fail seeded mismatches against a locked independent expectation set; human review evaluates the explanation as well as the machine result.
Supported learning environment
Approved local runtime · Microsoft Fabric / Power BI (approved lab only)
Scope and limitations
Use public or synthetic data. Keep customer evidence private. A course completion check does not qualify a live report or award a professional credential. Professional certification is planned and requires separate human award authority.
Curriculum source reviewed 2026-10-25. Durations are teaching estimates including exercises.