Meaning
The practice of designing, structuring, and optimizing inputs to elicit the most accurate and useful outputs from generative AI models.
Three examples from work
System instructions
DevThrough careful prompt engineering, we eliminated formatting errors in JSON responses.
LeadDid you include few-shot examples in the system prompt?
Model tuning
POShould we fine-tune an open-source model for tone of voice?
AI EngineerGood prompt engineering and persona framing are usually enough without training costs.
Quality assurance
QAHow do we test changes to our production prompts?
LeadWe run an automated evaluation dataset to score accuracy before releasing new prompts.
Techniques include role assignment, few-shot demonstration, chain-of-thought prompting, and structured output schema definitions.