A friend thinks you can answer this question about Streaming Data Pipeline
Your team is debating whether to build a Kafka + Flink streaming platform as the foundation for the data team. Most use cases are dashboards reviewed daily, with one fraud-detection model needing sub-minute features. What's the right architecture?