Cookable

Inspiration

We know how difficult it can be to go through all of your available ingredients accross your house, just to realize you are missing an ingredient or don't even know how to make any of the possible recipes, if you can even come up with some in the first place.

What it does

Our project enables you to speed up this process drastically, providing you with quick ingredient detection and generating a curated list of recipes based on your preferences and available ingredients: from pesketarian recipes to a quick 15 minute meal, our project can come up with recipes that are the right fit for you. Once you find a recipe you like, a quick tap will allow you to swipe through step by step instructions for how to make it, along with detailed images for each part of the process.

How we built it

We used Swift to develop our front-end app, providing a complete user interface. We wrote our backend in Python using the Flask library to streamline API development and processing. To properly manage and extract meaningful information from a variety of recipe and ingredient datasets, we used Palantir's AIP, enabling quick extraction of significant data, such as dietary restrictions or information within each recipe, and allowing us to standardize information accross a variety of datasets. To run our image-generation and classification models, we hosted our models on AMD's MI300X GPUs to support quick generation and processing.

Challenges we ran into

We struggled with using new technologies such as Palantir's AIP at first, but once we began to understand it its value and ability to simplify and improve our project became abundantly clear.

Accomplishments that we're proud of

We are proud of how we were able to develop a working product that provides consistent and effective outputs for real-world data. Additionally, we are proud of our ability to take advantage of new technology provided to us to support our development process, and how we were able to bring together a variety of components into one complete product.

What we learned

We learned a lot about Palantir's AIP and AMD's cloud compute services, along with how to take advantage of cloud based services and technologies along with our existing backend and frontend framework experience to develop a complete full stack application.