Description
NVIDIA is seeking a software engineer for its AI networking acceleration team to develop a highly optimized, open-source inference framework. The role involves building low-level infrastructure that runs on large supercomputers and data centers, using hardware offloads, GPU kernels, and RDMA network cards. Candidates need a B.Sc. or equivalent in Computer Science or Software Engineering, at least five years of modern C++/C/Python development experience, at least three years of Linux experience, and deep TCP/IP knowledge, with Linux internals, CUDA, ML frameworks, parallel programming, and high-performance computing experience preferred.
