The pipeline you saw in the previous section does not specify any jobs, as you may recall from the section on classic builds. Instead, it contains a list of tasks under the steps keyword. This means that it implicitly contains only a single job. With YAML pipelines, it is also possible to create a definition that contains more than one job. To do this, the following structure can be used:

trigger: 
- master

pool:
  name: Azure Pipelines
  vmImage: windows-2019

jobs:
- job: job1
  displayName: A pretty name for job1
  steps:
  - task: DotNetCoreCLI@2
    ...
- job: job2
  displayName: My second job
  pool:
    name: Azure Pipelines
    vmImage: ubuntu-18.04
  ...

Instead of adding the steps keyword directly to the pipeline, first, a list of jobs is created. Within that list, one or more job keywords are added, followed by the name for that job. Next to this technical name, a display name (displayName) can be specified for each job.

As the second job in this example shows, it is also possible to specify which agent pool to use per job. When no pool is specified for a job, the default pool specified at the top of the file is used.

By default, all the jobs in a pipeline run in parallel, but there are control options available to change this.