Lesson 1
Prompt design as specification
A production prompt is closer to executable policy than casual prose. Make inputs, constraints and output expectations explicit.
All learning tracks
Applied AI EngineeringPractical 90 min
Prompt Engineering & Prompt Testing
Treat prompts like production assets: version them, test them, compare them and defend them against regressions.
PromptingEvaluationRegression
Learning outcomes
Create measurable prompt tests
Build prompt regression cases
Separate prompt quality from model quality
Guided lessons
Learn the engineering thinking
Lesson 2
Golden datasets
Create representative examples with expected behaviour so prompt changes can be evaluated consistently.
Lesson 3
Prompt regression testing
Compare new prompt versions against known cases, edge cases and adversarial inputs before release.
Lesson 4
Scoring AI output
Combine exact checks, semantic criteria, rubric evaluation and human review depending on the workflow.
Lesson 5
Prompt injection awareness
Treat retrieved and user-supplied content as untrusted input when instructions and tool access are involved.
Hands-on practice
Do something with what you learned
Do not just tick these mentally. Write the query, create the test matrix, refactor the code, or document the decision. Practical evidence is the point.