As a shopper, I want to receive ML-based recommendations in the app, so that suggestions feel more personalized than the placeholder logic.
Acceptance criteria:
- Recommendations endpoint now calls the ML service instead of placeholder logic
- If model is unavailable or user data is sparse, a safe fallback is used
- Configuration allows switching between baseline, enhanced model, or placeholder if needed
Metadata
As a shopper, I want to receive ML-based recommendations in the app, so that suggestions feel more personalized than the placeholder logic.
Acceptance criteria:
Metadata
user story