Spam Email Detection Pipeline

Design a machine learning system that classifies incoming emails as spam or not spam.

Scenario: An email provider wants to filter spam for millions of users. The system must process incoming emails in real time and handle evolving spam techniques.

Your Task: Implement a function that returns the system design specification as a structured dictionary. Your design should include:

1. Data Pipeline - How to collect and preprocess email data
2. Feature Engineering - What features to extract from emails
3. Model - What ML model(s) to use for classification
4. Serving - How to serve predictions in real time
5. Monitoring - How to detect model degradation

Evaluation: Your design will be evaluated on completeness, appropriate model choices, and production readiness.