Prompt Engineering

Prompt engineering is the practice of designing the text you send to an LLM to get the output you want. It is more art than science — but with patterns that consistently work.

Key techniques: **clear instructions**, **examples (few-shot prompting)**, **chain-of-thought** (asking the model to reason step-by-step), **structured outputs** (asking for JSON), **role prompting** (telling the model who it is). Each can dramatically change output quality.

In production Gen AI engineering, prompt engineering is half of the engineering work. The other half is evaluation — knowing when your prompts are getting worse on edge cases as you iterate.