HPC & Me

Inspiration

One of the hardest things about working in high-performance computing is dealing with the complex scripts needed to submit jobs. In addition, errors resulting from jobs running can be quite intimidating. That's why we wanted to create an environment where working with HPC clusters is more accessible and intuitive.

What it does

It allows users to submit HPC job and command requests in natural language without needing to understand HPC schedulers, resource allocation, or shell scripting.

How we built it

The project is a full-stack web application using Next.js, Python/FastAPI, and OpenRouter for the AI integration. Cursor was used to accelerate development and debug errors. An iterative process was employed so that each feature could be effectively integrated.

Challenges we ran into

The backend proved to be highly complex to develop, integrate, and test.

Accomplishments that we're proud of

For our entire team, this was one of the most complex projects we've done. We were able to create a system that converted natural language to HPC jobs, supported real-time progress updates, and had a drag-and-drop feature for added interactivity.

What we learned

We gained an understanding of how HPC schedulers work and what they're capable of. We also learned of the importance of good UI/UX design, so that we will meet our users' needs.