Inspiration

Often times, developers have trouble finding the best model, package, or library to use for their project. We wanted to create not just a model finder, but a personalized one that intakes developers' experience levels as well as project in order to recommend them with the most personalized and popular models, packages, and libraries.

What it does

AI^2 Model Finder is an AI model that helps developers find the best AI model, package, or library.

How we built it

We used the AWS Bedrock Anthropic's Claude 3 Haiku Model to generate suggested AI model, package, or library, as well as GitHub links. We used Launch Darkly to implement A/B testing to provide users with explanations of different lengths to track user retention on our site. We used ConvexAI to connect our backend and frontend by allowing us to gather user input and return a response generated by our AWS Bedrock model.

Challenges we ran into

All of us were new to AWS Bedrock, Launch Darkly, and ConvexAI, so we had trouble connecting the frontend and the backend with the tech stack.

Accomplishments that we're proud of

We are proud of our vision for the product and have made the most important part of the product work that generates a suggested AI model, package, or library based on user input. We learned a lot more about these three tech applications we worked with!

What's next for AI^2 Model Finder

We want to finish connecting the tech stack so the frontend is fully connected to the backend. We also want to finish the UI/UX design.

Built With

Share this project:

Updates