`pyspark` connector to read & write from/to `Vertex AI Feature Store`

- Currently, one can only use ai platform sdk or pandas to do CRUD operations on Vertex AI feature store.
- It would be nice, if there's a way to utilize `pyspark` to do the same so that one can scale their ETL for these operations for large datasets

The issue names no files, tests, or entry points; start by reviewing the existing AI Platform SDK and pandas CRUD paths for Vertex AI Feature Store and how a PySpark ETL connector would fit. Done means a defined way to read and write Feature Store data through PySpark at large-dataset scale.

Written by the indexing model from the issue text.