Prompt engineering is the discipline of writing inputs that get reliable, high-quality output from an LLM. It covers techniques like role prompts ("You are a senior copywriter"), few-shot examples, structured output formats (JSON, XML), chain-of-thought hints, and constraint specification.
Good prompt engineering can change an LLM's output quality more than swapping models. The skill is increasingly automated by tools that test prompt variants against eval suites, but the underlying judgement of what to ask for remains human work.