← All courses

EDGE FDE

Advanced

Turn feedback into a better local model

Improve quality from real use without sending raw user text to a public trainer by default.

Signals to synthetic data to LoRA to a baked Core AI export, with honest privacy limits.

Collect Swift signals on device. Train and export in Python on a Mac. Load the bake in Swift. iOS 27 and macOS 27 only.

Lessons

  1. Feedback to a better local model
  2. Collect signals, not raw chats
  3. Turn signals into synthetic data
  4. Train a LoRA and export Core AI
  5. Honest privacy limits

Key concepts

Takeaways

Agent briefs

Related courses

LLM LoRA for your writing style · Test AI behavior before you ship · Version prompts like product copy · PyTorch to Core AI in Xcode · Core AI vs Core ML vs MLX

Official references

Apple Core AI · apple/coreai-models · Evaluations