Fine-Tuning: When to Use It

~10 min read

Fine-tuning is the right tool when you need to change the model's structure — its behavior, vocabulary, or writing style — rather than give it new knowledge. It changes HOW the model responds, not WHAT it knows.

Fine-Tuning: When to Use It is a Pro topic

Sign in, then upgrade to Pro or Power to unlock this topic and the full AI Engineering curriculum.

Key points

  • Fine-tuning changes the model's structure/behavior — vocabulary, style, reasoning patterns — not its knowledge
  • The book's example: internal meeting jargon needs consistent fluent handling, not a one-off lookup
  • Fits: consistent output format/tone across every response, domain-specific reasoning conventions, context too expensive to re-supply every prompt
  • Fine-tuning requires an actual training run — real compute cost, data curation, and time, unlike RAG's near-instant updates
  • Reaching for fine-tuning when the actual gap is missing knowledge (RAG's territory) pays this cost for the wrong problem