Ruskaruma

I focus on how models and agents actually run, building high-performance AI runtimes: inference engines, CUDA kernels, and agent execution layers.

Alongside deep work in compiler design, I'm an active contributor to open-source AI systems and infrastructure.