MealPrepAI

Inspiration

We were inspired by the daily struggle our friends, fellow students, and young professionals face: the desire to eat healthy, home-cooked meals versus the reality of a busy schedule. The mental load of planning meals, figuring out what to buy, and learning how to cook is a significant barrier. We saw an opportunity to use AI to create a truly personal and intelligent kitchen assistant that removes the friction from healthy eating, making it accessible and enjoyable for everyone, regardless of their time or cooking experience.

What it does

MealPrepAI acts as a personal AI chef in your pocket. At its core, it learns your dietary goals, preferences, and allergies to generate tailored meal suggestions. Its standout feature is its ability to create delicious recipes based on the ingredients you already have in your pantry, minimizing food waste and saving you a trip to the store. It also generates a smart grocery list for the items you're missing and guides you through the cooking process with timed, step-by-step instructions, making cooking less intimidating and more fun.

How we built it

This project is a full-stack application built with a modern, scalable architecture to support our team's concurrent development.

Frontend: We used React to build. We focused on a clean, component-based architecture with React Navigation for routing and the Context API for state management.

Backend: The backend is a powerful and modular API built with Django and PostgreSQL. We intentionally separated concerns into distinct apps for users (authentication) and api (core business logic) to ensure maintainability and allow for parallel development.

AI & Machine Learning: The "brains" of our app is a large language model fine-tuned on AWS Bedrock. We curated a high-quality dataset of recipes to train the model to output consistently structured, timed, and nutritionally-aware JSON data.

Challenges we ran into

Accomplishments that we're proud of

We are incredibly proud of building a functional, end-to-end AI application that solves a real-world problem. Our greatest accomplishment is the successful fine-tuning of our custom model on AWS Bedrock, which now serves as a reliable and intelligent core for our recipe generation. We're also proud of the clean, well-documented full-stack architecture we designed, which has been instrumental in our team's ability to collaborate effectively. Seeing the "pantry-to-plate" feature come to life—where the app generates a great recipe from a random list of ingredients—was a huge moment for us.

What we learned

This project was a tremendous learning experience. We learned the critical importance of a well-defined API contract between the frontend and backend before writing a single line of code. We gained hands-on experience with the entire MLOps lifecycle, from data curation and model fine-tuning to deploying and calling a custom AI model via a REST API. Most importantly, we learned how to effectively organize a collaborative software project on GitHub using a structured branching strategy, issue tracking, and clear documentation.

What's next for MealPrepAI

We're just getting started. Our roadmap includes expanding our grocery store integrations to more vendors to give users more choices. We plan to introduce social features, allowing users to share their favorite recipes and weekly meal plans with friends.