For a one-off requirement the path of least resistance I feel would be
load it into a local copy of Oracle installed on your local machine using the external table method.
Take a datapump export of the table you loaded in
Use the s3_integration option to use datapump instead to load in your data as described here - https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/Oracle.Procedural.Importing.DataPump.html
For a repeating requirement I would use DMS serverless here
Load your CSV file(s) into a S3 bucket
Configure your solution using Terraform or another IAC method to easily reproduce the config as needed
Use the DMS serverless option to help reduce the operational overhead associated with a DMS configuration https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Serverless.html
DMS does support S3 as a source https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Source.S3.html and Oracle as a target https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Target.Oracle.html