TECHNICAL / PyTorch internals
Inside torch.compile: From Python to Kernels
The compiler architecture first: graph capture, optimization, and runtime reuse. Then follow a small tensor function through the stack.
LYNN / ENGINEERING & NOTES
Systems, in detail.
Life, in between.
I'm Lynn, an AI infrastructure engineer.
I write about how systems work, where they break,
and what it takes to make them better.
Writing
A WORK IN PROGRESSNo posts here yet. More to come.
MAIN THREAD
Under the hood.
LLM training and inference, reinforcement learning infrastructure, GPU performance, and the engineering decisions behind reliable AI systems.
Technical writing