Senior Deep Learning Software Engineer, Inference

Role overview Build and optimize software that serves deep learning models efficiently across GPU platforms, from data center accelerators to edge systems. The work centers on inference performance for large language, multimodal, and generative AI models, with contributions to open-source frameworks and production serving solutions.

Responsibilities - Profile, analyze, and tune deep learning models to improve inference performance. - Scale model performance across accelerator architectures and deployment environments. - Develop features and contribute code to inference libraries and projects such as vLLM, SGLang, and FlashInfer. - Apply GPU kernels and related tools to improve model serving pipelines. - Coordinate with framework and library teams to deliver inference optimizations. - Engage with the deep learning community to bring current algorithms into public software.

Requirements - A master’s or doctoral degree in a relevant field, or equivalent experience. - At least five years of relevant software development experience. - Strong C and C++ programming and software design skills. - Interest or experience in optimizing deep learning inference and working with high-performance software.

Nice to have - Python experience and familiarity with agile software development. - Production experience training, deploying, or optimizing deep learning models. - Performance modeling, profiling, debugging, code optimization, or CPU/GPU architecture knowledge. - GPU programming with CUDA, Triton, or CUTLASS; contributions to projects such as PyTorch, vLLM, or SGLang; or multi-GPU communications experience with NCCL or NVSHMEM.

Benefits and work setup The listing describes a hybrid arrangement and includes equity and benefits. Stated base salary ranges are $152,000–$241,500 for one level and $184,000–$287,500 for another; pay depends on location and experience.