✨ Prompt Engineering Mastery

# Give the AI a specific role
prompt = """
You are an experienced Python developer with expertise in data science.
Help me optimize this code for processing large datasets:

[code here]

Provide specific suggestions for improving performance.
"""

# Works well for:
- Expert advice
- Technical writing
- Specialized knowledge

# Provide examples of desired output
prompt = """
Extract the company name and job title from these texts:

Text: John Smith works as a Data Scientist at Google
Output: {"company": "Google", "title": "Data Scientist"}

Text: Sarah is the CTO of Microsoft
Output: {"company": "Microsoft", "title": "CTO"}

Text: Alex Martinez is employed by Tesla as a Software Engineer
Output:"""

# AI learns pattern and applies to new example

# Encourage step-by-step reasoning
prompt = """
Problem: A store has 23 apples. They sell 17 in the morning and receive 
a delivery of 42 more. How many apples do they have now?

Let's solve this step by step:
1. Start with: 23 apples
2. After morning sales: 23 - 17 = 6 apples
3. After delivery: 6 + 42 = 48 apples

Therefore, they have 48 apples.

Now solve this: A library has 156 books. They lend out 89 and purchase 67 new ones.
How many books do they have now? Let's solve step by step:
"""

# Dramatically improves reasoning accuracy!

# Request specific format
prompt = """
Analyze the sentiment of this review. Return JSON with score and reasoning.

Review: "The product works okay but shipping was slow."

Format:
{
  "sentiment": "positive/negative/neutral",
  "score": -1 to 1,
  "reasoning": "explanation",
  "key_aspects": ["list", "of", "aspects"]
}
"""

# Gets consistently formatted outputs

import openai

prompt = "What is 1547 + 3892? Show your work."

# Generate 5 different responses
responses = []
for i in range(5):
    response = openai.ChatCompletion.create(
        model="gpt-4",
        messages=[{"role": "user", "content": prompt}],
        temperature=0.7
    )
    responses.append(response.choices[0].message.content)

# Find most common answer
# More reliable than single response!

# Provide relevant context from your data
context = """
[Retrieved relevant documents from your database]
"""

prompt = f"""
Based on the following context, answer the question.

Context:
{context}

Question: What is our return policy?

Answer based only on the provided context. If the answer isn't in the context, say so.
"""

prompt = """
Problem: Plan a 5-day trip to Japan

Generate 3 different approach strategies:

Strategy A: [Focus on culture and temples]
Strategy B: [Focus on food and cuisine]
Strategy C: [Focus on nature and hiking]

Evaluate each strategy:
Strategy A evaluation: [pros and cons]
Strategy B evaluation: [pros and cons]
Strategy C evaluation: [pros and cons]

Select best strategy: [choice with reasoning]

Final detailed plan based on chosen strategy:
[day by day itinerary]
"""

"""
Write a Python function that:
- Takes a list of integers
- Filters out even numbers
- Returns the square of odd numbers
- Include docstring and type hints
- Add example usage
"""

"""
Write a LinkedIn post about AI in healthcare.

Requirements:
- Tone: Professional but conversational
- Length: 200 words
- Include: 1 statistic, 1 real-world example
- Add: 3 relevant hashtags
- Hook: Start with a surprising fact
"""

"""
Analyze this sales data and provide insights:

[CSV data here]

Generate:
1. Top 3 insights (with numbers)
2. Trends over time
3. Recommendations (specific actions)
4. Data visualization suggestions

Format as markdown with clear headers.
"""

"""
You are a customer support agent for TechCo.

Guidelines:
- Be empathetic and professional
- Offer specific solutions
- If you can't help, escalate politely
- Keep responses under 100 words

Customer message: [message]

Response:
"""