Prompt Eng Best Practice

#### **1. Set the stage (Context):**

- Tell the AI who it's talking to.
- Define the expertise level needed.
- Specify any constraints or limitations.

###### **Example:**

- **Bad Prompt**:
  > Analyze this dataset
- **Meh Prompt**:
  > Analyze sales trends in this CSV
- **Perfecto Prompt**:
  > You are a data specialist analyzing our quarterly sales data. Identify key revenue patterns, create visualizations, and highlight anomalies, focus on customer segments and regional variations.

---

#### **2. Define the task (Action):**

- State exactly what you want.
- Include specific requirements.
- Break complex tasks into steps.

###### **Example:**

- **Bad Prompt**:
  > Help me learn python
- **Meh Prompt**:
  > Teach me python data structures
- **Perfecto Prompt**:
  > You are a senior python instructor. Create a step by step learning plan for data structures, including:
  >
  > - Practice exercises for lists and dictionaries.
  > - Real world examples using (Pandas)
  > - Common pitfalls to avoid.

---

#### **3. Specify the Output (Format):**

- Define the structure.
- Clarify the Tone & Style.
- Set length.

###### **Example:**

- **Bad Prompt**:
  > Make a content plan
- **Meh Prompt**:
  > Create a formatted content plan with specific sections
- **Perfecto Prompt**:
  > Present the plan in a table format with columns for platform, content type and key message. Include engagement hooks posts under 200 characters

---

#### **4. Use Role Assignment:**

###### **Example:**

- **Bad Prompt**:
  > Debug this code
- **Meh Prompt**:
  > Debug this python Machine Learning Model
- **Perfecto Prompt**:
  > As a senior ML Engineer specializing in Neural Networks:
  >
  > - Review the PyTorck implementations
  > - Identify performance bottlenecks
  > - Suggest optimization strategies for training speed

---

#### **5. Add specific Constraints:**

###### **Example:**

- **Bad Prompt**:
  > Analyze this dataset
- **Meh Prompt**:
  > Analyze customer churn patterns
- **Perfecto Prompt**:
  > Analyze our customer churn data set with these parameters
  >
  > - Focus on inactive users > 30 days
  > - Compare behavioral patterns pre-churn
  > - Include statistical significance tests
  > - Output visualizations in seaborn

---

#### **6. Request Examples & Variations:**

###### **Example:**

- **Bad Prompt**:
  > How do I implement this algorithm
- **Meh Prompt**:
  > Show different ways to implement this algorithm
- **Perfecto Prompt**:
  > Provide 2 implementations of this sorting algorithm:
  >
  > - Optimized for memory efficiency
  > - Optimized for processing speed, include time complexity analysis for each implementation