Gen AI — Prompt Engineering

What is prompt engineering?

Designing inputs to get desired outputs from AI models.

Basic principles

1. Be specific: Clear instructions
2. Provide context: Background information
3. Set constraints: Output format, length
4. Use examples: Show desired output

Bad vs good prompts

# Bad
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Write about dogs"}]
)

# Good
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": """
Write a 100-word informative paragraph about golden retrievers.
Include their temperament, exercise needs, and suitability for families.
Use a friendly, accessible tone.
"""}]
)

Prompt templates

def create_prompt(topic, audience, length):
    return f"""
Write a {length}-word explanation about {topic}.
Target audience: {audience}.
Tone: Professional but accessible.
Include one practical example.
"""

Role prompting

response = client.chat.completions.create(
    model="gpt-4",
    messages=[
        {"role": "system", "content": "You are an expert Python developer."},
        {"role": "user", "content": "Write a function to sort a list"}
    ]
)

Output formatting

response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": """
List 3 Python libraries for web scraping.
Format as:
1. Library name - Brief description
2. Library name - Brief description
3. Library name - Brief description
"""}]
)

Mini Practice

1. Rewrite a bad prompt to be specific
2. Create a prompt template
3. Use role prompting
4. Specify output format

Up Next

Continue with Zero-Shot - No examples needed.