Seems to me that the issue is multiple files of CSV that should be grouped as one ?
If that’s the case, then perhaps the system needs to have a mechanism to keep track of multiple files as a single upload. i.e. metadata. This way, the files could be retrieved easily from S3 as a group.
Depends on your use case, architecture wise you could try to separate the web app server and the processing server. Each uploaded files will be queued to be processed (parsed).
Regards,
Wan.