Evaluating Frontier Models for Performance and Representation
Our project is part of the Break Through Tech AI Program, where we work in collaboration with Relativity to tackle the issue of bias in AI models. We are developing tools and methodologies to analyze models, uncover inherent biases, and explore strategies to reduce or eliminate these biases.
- Python
- Pandas, NumPy, Langtest
Equipment Scheduler
We are developing an equipment scheduler using React and JSX, powered by a SQL database. Users can select a school and log in via a token sent to their email. Once logged in, they can schedule equipment, bookmark items, and staff can add new equipment. The project emphasizes real-time availability, streamlining the scheduling process for an improved user experience.
- React
- SQL
- HTML
Sade Song Searcher
Developed a dynamic webpage using Bootstrap to showcase Sade’s discography, featuring an interactive interface for exploring albums, viewing song details, and implementing a search feature for finding songs by partial lyrics, while collaborating on backend data integration for seamless database functionality.
- HTML, CSS, JS
- EJS, SQL
Travel Journal
Created a comprehensive travel planning application using Java and Android Studio, integrating Firebase for data storage and enabling users to create travel logs, plan trips, and research destinations. Features includes a search function, journal entry management (CRUD)
- Java
- Firebase
- Android Studio