The process of integrating a machine learning model into a production environment, making it accessible for end-users or other systems to generate predictions or insights

Model Deployment Modes

<aside> 💡

Comparing methods

</aside>

Why is Spark Slow in Online Scoring?

<aside> 💡

Spark is primarily designed for batch and streaming deployments

</aside>

Deploying a Single-Node model on a Spark DataFrame

Deploying a Spark ML Model on a Spark DataFrame

Code Example

Conclusions:

• Spark ML is good for batch/streaming but not online stuff.
• Can do distributed computing for data processing, training.

Inference with Spark

Approaches to Model Inferencing with Spark

<aside> 💡

Methods for Deploying and Running Inference

</aside>