📋 Prompt Engineering Cheat Sheet

Essential prompt engineering techniques for the AIF-C01 exam.

Prompting Techniques

  • Zero-shot: ask the model directly without examples.
  • One-shot: provide one example before the actual query.
  • Few-shot: provide multiple examples to guide the model.
  • Chain-of-thought: ask the model to reason step by step.
  • System prompts: set the model's persona, constraints, and behavior.

Best Practices

  • Be specific and clear — vague prompts produce vague outputs.
  • Provide context and constraints to narrow the response.
  • Use delimiters (###, ```) to separate instructions from content.
  • Specify the output format (JSON, bullet points, table).
  • Iterate and refine — prompt engineering is an iterative process.
  • Test with diverse inputs to ensure robustness.

Common Pitfalls

  • Hallucinations: model generates plausible but incorrect information.
  • Prompt injection: malicious inputs that override system instructions.
  • Context window limits: too much input causes truncation.
  • Over-reliance on temperature: too high = nonsense, too low = repetitive.
  • Not validating outputs: always verify critical AI-generated content.

When to Use What

  • Zero-shot for simple, well-defined tasks.
  • Few-shot when the model needs examples of the desired format.
  • Chain-of-thought for complex reasoning or math problems.
  • System prompts to enforce consistent behavior across conversations.
  • RAG when the model needs access to current or proprietary data.

Practice Prompt Engineering Questions

Put your knowledge to the test with practice questions.

More AIF-C01 Cheat Sheets