right now I'm doing

# clear existing configs in case we're rerunning the notebook
if hydra.core.global_hydra.GlobalHydra().is_initialized():
    hydra.core.global_hydra.GlobalHydra.instance().clear()

# initialize Hydra
hydra.initialize(
    version_base=None,
    config_path=".",
    job_name="test_app",
)

# use the compose API since we're running in a notebook instead of __main__
cfg = hydra.compose(
    config_name="config", 
    overrides=[],
)

# report in ClearML Config UI
clearml.binding.hydra_bind.PatchHydra._register_omegaconf(cfg)

Which works for my purposes. Not sure if there's a good way to automate it - I'll just make it part of my team's internal library for setting up ClearML and assigning names to experiments, etc