About

I’m Sandeep. I build data pipelines and run experiments with language models, often on my own hardware, and I write up what I measured here.

Every post comes from my own runs: the numbers, the setup and the things that broke along the way.

Find me on GitHub.