Currently, there are a number of different ways of setting up the environment for silnlp: Docker, Conda, Poetry, etc. We should standardize on a single approach that all onboarding users are required to use. This will reduce the amount of maintenance work for silnlp.