ML Model Selection Advisor — Choose the Right Algorithm

You are a senior ML engineer. Help me choose the right model for my problem.

**My problem:**
- Task type: [classification / regression / clustering / ranking / anomaly detection / time series / NLP / recommendation]
- Dataset size: [rows] rows × [columns] features
- Target variable: [describe what you're predicting]
- Feature types: [numeric / categorical / text / mixed]
- Data quality: [clean / some missing / messy]

**Constraints:**
- Interpretability requirement: [must explain predictions / black box OK]
- Latency requirement: [real-time <100ms / batch OK]
- Training compute: [laptop / single GPU / cloud cluster]

**Your analysis should include:**
1. **Top 3 recommended algorithms** ranked, with:
   - Why it fits this problem
   - Expected performance range
   - Training time estimate
   - Pros and cons for this specific case
2. **Baseline model**: The simplest model to try first (and expected accuracy)
3. **Feature engineering suggestions**: Top 5 transformations likely to improve performance
4. **Evaluation strategy**: Which metrics to use (and which NOT to use, and why)
5. **Common pitfalls** for this problem type
6. **Scaling path**: What to try if the first model isn't good enough

**Do NOT recommend deep learning unless the dataset has >100K rows or involves images/text.**