Now

I'm currently deep in the CUDA programming model — writing kernels, understanding warp-level execution, and slowly building intuition for how data movement dominates everything on the GPU. It's humbling in the best way.

On the application layer, I'm building a multi-agent CMS that coordinates several LLM-backed specialists through a shared context layer. The goal is to let agents plan, edit, and publish content collaboratively, with a human in the loop only when it matters.

I'm also working through Programming Massively Parallel Processors (4th edition) cover to cover. The exercises are brutal. Highly recommended.

When I'm not heads-down in code, I'm listening to lo-fi streams, tinkering with Linux on unconventional hardware, and trying to understand rotational dynamics well enough to explain it to someone else.