Inspiration

The graphics card market is constantly changing — dozens of models, confusing specifications, and widely varying prices make it difficult to know what to choose.

What it does

GPYOU simplifies that process by asking you a few key questions, then offering personalized recommendations tailored to your situation.

How we built it

We fed a dataset of GPUs and their specifications into Gemini 2.0 Flash using Gemini API and pseudo-trained it on this data, then by getting answers for specific questions from the user and using their answer in a prompt, we are able to effectively use Google Gen Ai to get a personally tailored GPU recommendation based on the user's needs.

Challenges we ran into

We ran into many challenges, the biggest of which was connecting the javascript that has the GenAi with our html/css website. We ran into problems as we were trying to make use of two seperate html files, one for the user's input page and one for the recommendation page, however, this plan was flawed as using the same script for two html files was inefficient. We fixed this problem by having two divisions that act essentially as two pages. Once the input is taken and a recommendation is given, the input page is hidden and the recommendation page start showing.

Accomplishments that we're proud of

We are very proud that we were able to integrate 3D models of most GPUs in our list to include in the recommendation page, as we thought this was essential for anyone who wants to buy a GPU, afterall, if you don't like how it looks, then chances are you are not going to like it.

What we learned

As a team, none of us had any prior knowledge in using and integrating a generative AI into a website, and it was truly a fun experience to learn how to. Especially with Gemini's ease of usage, it didn't take us a long time to set it up, make a dataset, and construct the perfect prompt for the perfect recommendation.

What's next for GPYOU

Now that we have a solid foundation for GPYOU that helps many people who do not necessarily have the time to do countless hours of research to pick a pc part, we can easily expand to include CPU, Motherboard, RAM and other pc components in our recommendations.

Share this project:

Updates