Prompt Engineer
Improve a prompt while keeping its evaluation and release history clear.
Prepare reviewable inputs for an AI-assisted workflow
Run the mortgage kit and inspect the income discrepancy. Define which source fields an assistant may use, which statements require evidence, and where a human must review its recommendation.
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
Notes and progress are saved on this device.
Make every prompt change reviewable.
Improve a prompt while keeping its evaluation and release history clear.
Find an existing prompt or create a candidate with an explicit task and input structure.
Build a representative evaluation set and agree what a useful and acceptable response looks like.
Compare prompt versions against the same cases and document failures, quality, and cost differences.
Submit the selected version for review and record the released version for application teams.
Expected handoff
A reviewed prompt version with evaluation cases and release notes.
Explore your platform capabilitiesMeasure your progress
- Prompt changes supported by repeatable evaluations
- Time to identify the version behind an AI response
Use AI to suggest prompt variants and adversarial cases; independently review the evaluation criteria and outputs.