Deep Learning Kernel Optimizer

Looking for training scalable LLM / GPU Deep Learning Kernel optimization using Triton lang/TorchDynamo / model serving using Triton Server / HPC cluster setup to support RLHF.