The internship
I expected to work with a GPU during my first internship. I wanted to train a model and build a full platform around that work. I couldn't get the compute I needed, and the idea stalled.
I wanted to build. I couldn't get the GPU time.
I'm a student. This started with a project I wanted to take further during my first internship.
Where I got stuck.
The workaround
I already knew Google Colab. It was useful for smaller experiments, but the resources I could use there did not get me far enough with the larger project I had in mind.
The harder route
I found IndiaAI Compute, which includes students. But registration, verification, and a project request felt like a lot to navigate when I just wanted to test an idea. From where I stood, it seemed easier to find a route if you already had a team or institution behind you.
I'm the student behind YuvaCompute.
I started working on YuvaCompute after my internship project stalled. I'm figuring out how to get GPU time to students without making them write a long proposal just to begin.
From student ID to saved work.
A proposed path for students whose projects need more than their own laptop.
1Verify student
One short student ID check. The verification method still needs to be designed.
2Write code
Start with a browser notebook or workspace for your project.
3Run on a GPU
Use GPU time for demanding work. Supply, limits, and price are still open.
4Save and return
Keep code, files, and checkpoints for another session.
This is the experience I want to build. Student verification and GPU access are not available yet.