I am using Databricks community edition for running spark workloads . I understand it uses kernels to run the notebooks
Regarding the first question, Databricks has a limited number of supported "main" languages - Scala, Python, R, and SQL - you can set them as primary language when creating the notebook. Besides language set on the notebook level, you can use another language for a given cell by using magics, like, %scala, %python, %r, %sql. There are also additional magics, for example, %sh for execution of the shell code on the driver, %fs for working with files on DBFS, etc. All of this you can find in the documentation.
Regarding the 2nd question - actual implementation is not public, but it should work similarly to the Spark implementations - pyspark, etc.
