Comment by DebtDeflation

Fine tuning for training the model to perform a new task, RAG for adding knowledge.

In your example, you would fine tune the model to train it to code in a language it hasn't seen before, RAG will not really help with that.