Software Engineering Intern, DLFW Comms - 2027
Core
Integrate advanced communication technologies into AI stacks and develop high-performance kernels for multi-GPU scaling in Deep Learning applications.
Role type
Deep Learning Engineer (Communication & Kernels)
Builds
AI frameworks (PyTorch, vLLM, SGLang, TRT-LLM), communication libraries (NCCL, NVSHMEM), and custom compute-communication kernels.
Domain
Artificial Intelligence / High-Performance Computing / GPU Systems
Deliverable
production ML models
Required skills
CUDA kernel authoring, Python, C++, LLM parallelism understanding, AI training/inference background, rapid prototyping
Preferred skills
PyTorch, JAX, TRT-LLM, vLLM, SGLang, veRL, Expert Parallelism, Tensor Parallelism, Data Parallelism, Pipeline Parallelism, CUDA profiling, large-scale training stack
Technologies
PyTorch, vLLM, SGLang, TRT-LLM, veRL, NCCL, NVSHMEM, Triton, cuTe, CUDA, Python, C++
