Prompt Engineering

Prompt engineering is the practice of designing and refining the instructions (prompts) to give an AI model so it produces the desired output. A prompt can include:

instructions
questions
examples
constraints
desired output formats

Prompt engineering focuses on how to phrase a request, what structure to use, and what examples to include to guide the model’s response. LLMs respond differently depending on how a request is written. Prompt engineering is about crafting prompts that lead the model toward accurate, useful, and consistent results.

About Prompts

A well-designed prompt often includes several elements.

Clear instructions (tell the model what role it should take or what task it should perform)
- Example: “You are a technical writer. Explain the following concept for beginners.”
Context (provide relevant information the model should consider)
- Example: “This documentation is for developers who are new to REST APIs.”
Input data (the content the model should work on)
- Example: “Code snippet: function add(a, b) { return a + b; }”
Output format (specify how the response should be structured)
- Example: “Respond using the following format: Description: … Parameters: … Example: …”
Examples (show the model examples of the desired behavior)
- Example: “Input: HTTP 404 | Output: The requested resource was not found on the server.”

Why Prompt Engineering Matters

Effective prompts can:

Improve answer accuracy
Reduce hallucinations
Control tone and style
Produce structured outputs
Make AI behavior more predictable

Example

Basic prompt:

Explain recursion.

Improved prompt:

Explain recursion to a beginner programmer.  
  
Requirements:  
- Use simple language  
- Provide a short code example in Python  
- Limit the explanation to 150 words

Common Prompt Engineering Techniques

Zero-shot prompting

Ask the model to perform a task without examples.

Translate this sentence into Spanish.

Few-shot prompting

Provide examples to guide the model.

English: Hello  
Spanish: Hola  
  
English: Goodbye  
Spanish:

Chain-of-thought prompting

Encourage the model to show intermediate reasoning steps.

Solve this math problem step by step.

Role prompting

Assign the model a role to influence tone and expertise.

You are an experienced software architect.