Happy Plate

Over 20 million people in Germany suffer from the symptoms of food intolerances. The only way to reduce symptoms effectively is to reduce or omit the causing ingredients in their daily diet. People who have just unveiled their food intolerance have a hard time changing their diet because they most often eat or cook the same dishes that they have been used to all their lives. The project should support people by providing recipes that are safe to eat but also relate to their previous diet. The personalization of the recipes should increase the adherence rate of the new diet and thus help reduce the time to change their habits.
We think that modern LLMs can deal with this task in a reproducible manner. The goal should be, that users can interact with a UI that allows the user to input his preferences and that results can be tuned by further actions (e.g. when the user does not like eggplant, a new recipe without eggplant can be generated). The interaction however should not be in a prompt/chatbot-style, but the underlying prompt should be generated from the user’s properties.

Development Team: Megh Vora, Maria Kharlamova, Ali Karimiesfahani, David López Caballero, Veit Goetz, (left to right)

Scrum Master: Rebecca Lennartz, Charlotte Pradel

YouTube Link:

Partner: BesserEsser