179 questions
No questions match those filters.
← All questions
RAGEasy· Fine-tuning· Retrieval· Data Freshness
Would you fine-tune an assistant on a new price list to...
This is one of the questions in the full AI/ML interview bank. Pro unlocks all 1789 questions; Premium includes the same bank plus the highest daily Practice limit.
See plansA strong answer rejects the premise of fine-tuning for dynamic data like pricing. Fine-tuning a 27B model requires significant GPU resources (approx. 1,324 GPU-hours) and becomes obsolete as soon as the next price change occurs. In contrast, the same investment in compute can support roughly 15 million retrieval-augmented queries, making RAG the superior architectural choice for frequently updated information.