Hi ReassuredTiger98
I think you should have something like:

` @PipelineDecorator.component(task_type=TaskTypes.application, docker='clara_docker_container_if_we_need')
def step_one(param):
print('step_one')
import os
os.system('run me clara')
# I'm assuming we should wait?
return

@PipelineDecorator.component(task_type=TaskTypes.training)
def step_two(param):
print('step_two')
import something
somthing.to_do()
return

@PipelineDecorator.pipeline(name='custom pipeline logic', project='examples', version='0.0.5')
def executing_pipeline():
print('pipeline logic strating')

print('launch step one')
step_one('clara args?!')

print('launch step two')
step_two('do something)

if name == 'main':
PipelineDecorator.set_default_execution_queue('default')
# Start the pipeline execution logic.
executing_pipeline()

print('process completed') `Do we need to wait for clara to spin for step two ?

Notice the entire code (including two functions is stored on the pipeline Task itself, the step itself (i.e. the function) is then package as its own standalone Task, including reference to the docker image if needed, and required python packages based on what's imported inside the function)