Glossary
What is Prompt Engineering?
Prompt engineering is the practice of designing and refining the instructions given to a language model to reliably produce accurate, useful, and appropriately formatted output for a specific task.
The same underlying model can produce noticeably different quality output depending entirely on how a request is framed — how much context it's given, whether it's shown examples of the desired output, whether instructions are specific and unambiguous, and whether it's asked to reason step by step before answering. Prompt engineering is the discipline of deliberately shaping those variables instead of leaving them to whatever phrasing comes to mind first.
Common techniques include few-shot prompting (showing the model one or more examples of the input-output pattern desired before giving it the real task), chain-of-thought prompting (explicitly asking the model to reason through steps before giving a final answer, which measurably improves accuracy on multi-step problems), and structured output constraints (specifying an exact format — JSON, a particular template — so the response can be reliably parsed by other software downstream).
In a production system, prompt engineering is rarely a one-time task — prompts typically get versioned and tested against real examples much like code, since a small wording change can shift output quality in either direction, and what works well for one model can behave differently on another.