The filesystem you will work with is mainly split into:
Home directory for your user
Synchronized throughout all nodes (i.e., changes made here always take effect on all nodes)
Used for configuration files, environments, ssh config, ...
local scratch (i.e., not synchronized between nodes)
used for high-bandwidth needs only (e.g., ML dataset loading that would otherwise be a bottleneck if deployed to /work)
main directory for storing or loading work data
BeeGFS => parallel filesystem that saves on all available nodes (thus data stored here is available on any node under /work)