The problem
People make several connected decisions around food: what is already in the kitchen, what to cook, how to organize the week, and what to buy. Culinaire brings those decisions into a mobile product instead of making an AI conversation the entire experience.
My role
I work across product development and Flutter application engineering, connecting the cooking interface to recipe data, AI-assisted generation, account services, and planning workflows. The application and its companion Next.js marketing site are part of my consumer product work.
The implementation
The Flutter codebase connects Firebase account and data services, AI generation, recipe search, and Open Food Facts. Its meal-planning flow converts preferences into meal slots, retrieves recipe candidates, and supports grocery-list creation. Barcode scanning exposes available product information with pass, caution, and unknown allergen signals. The app also includes focused cooking steps rather than presenting every response as unstructured chat.
Important decisions
- Use structured recipe and planning data so generated suggestions become useful interface actions.
- Keep dietary preferences connected to the user's cooking workflow.
- Represent incomplete food information as unknown.
- Connect subscription access and usage controls to the features they govern.
Live product
Culinaire has public Android and iOS listings and an official product website. The current codebase continues to evolve. Its food signals are informational; this case study makes no medical or allergy-safety guarantee and uses no unsupported adoption metrics.
The system view illustrates the relationship between preferences, recipes, planning, and groceries. The public links provide a direct view of the released experience.
