Hendrik Sippel

Building an AI-powered Meal-Planning Engine for Elderly Homes

Nobody thinks of menu planning as an AI problem until you run a kitchen for a nursing home. Dozens of dietary profiles, allergen requirements, rotating product availability, tight inventory and no visibility into budgets, replanned every single week. At Pyne, we partnered with a German start-up to turn that domain expertise into a working product for elderly care facilities. The result is an AI-powered engine consisting of three layers: an agentic recipe generation, a menu optimization algorithm and a platform that ties it all together with inventory and supplier data. 

In this talk, Hendrik walks through the architecture, the tradeoffs at each layer and what it actually took to get from chicken nugget suggestions to well-balanced, nutritious meals. His talk shows how different AI techniques combine to solve a real-world optimization problem.

Bio: I’m a Data & AI Engineer at Pyne where our team tackles technically complex challenges every day. With a background in Data Science, I have gained extensive experience in training, deploying and embedding AI into products and services people use every day.