I started working at NVIDIA TensorRT-LLM. I will research how to automate inference optimization across kernels and graph compiler layer.