GPU Engineer

Kog is seeking a GPU Engineer to optimize real-time LLM inference for AI agents by co-designing model architectures, inference engines, and low-level GPU kernels. The role involves developing and optimizing GPU kernels using C++, CUDA, HIP, and low-level instructions to minimize generation latency on NVIDIA and AMD datacenter GPUs. You will also help build internal AI agents to automate and accelerate the implementation and optimization of new AI models.