Integrating Spark ML Pipeline into a Highly Loaded Low-Latency Recommendation Service

The speaker will describe how you can implement a Spark ML Pipeline trained on big data on a cluster into an online recommendation service that has fairly stringent response time requirements.