Core Techniques for Prompt Engineerings
Core Techniques are the prompting methods that reliably improve results. They include zero-shot and few-shot prompting, assigning a role, stating constraints, and giving explicit success criteria. Clear instructions beat clever tricks: say who the audience is, what to do, what to avoid and what the answer should look like.
Stage: Foundations · Part of the Prompt Engineering roadmap · Last reviewed 19 November 2026
What to learn in Core Techniques
- Add two or three few-shot examples that show the exact output format
- Rewrite a vague instruction into specific, testable requirements
- Use delimiters to separate instructions from supplied text
- Specify what to do instead of only listing what to avoid
- Iterate on a prompt by changing one thing at a time
Core Techniques interview questions
When does few-shot prompting beat zero-shot prompting?
Few-shot helps when the task has a particular style, label set or output format that is hard to describe in words. The examples show the pattern directly. For simple, well-known tasks, a clear zero-shot instruction is often enough and uses fewer tokens.
Why use delimiters in a prompt?
Delimiters such as triple quotes or XML-style tags mark where instructions end and the supplied data begins. This reduces confusion, makes parsing easier, and lowers the chance that text inside a document is mistaken for a command.
Where Core Techniques fits in the Prompt Engineering roadmap
Core Techniques belongs to the Foundations stage (beginner level), which takes about 12 hours. How LLMs respond.
Other units in this stage
Back to the full Prompt Engineering roadmap