Taste inference + pairing logic

Flavor Compass

Enter your top dishes and get recommendations that balance fit and discovery. This is a transparent recommendation engine prototype focused on product decision quality: clear inputs, controllable risk, and explicit reasoning.

Chef output

Suggested Pairings

Press Generate Picks to get recommendations.

Why This Demonstrates Product Thinking

Input quality: asks for concrete user signals (top dishes), not abstract survey ratings.
Control surface: novelty slider gives users explicit control over exploration risk.
Explainability: every recommendation includes fit score, novelty score, and reasoning.
Operational posture: can be instrumented with second-use acceptance as the core quality metric.