Senior Deep Learning Framework Communications Engineer
$184k - $357k • Remote • US, CA, Santa Clara
Posted 6d ago
About the job
NVIDIA is seeking a motivated Deep Learning engineer to integrate advanced communication technologies into AI stacks, including PyTorch, TRT-LLM, vLLM, and SGLang. You will work with the team that created communication libraries like NCCL and NVSHMEM, and technology like GPUDirect, to scale Deep Learning and HPC applications. Your work will directly impact AI application performance by improving communication between GPUs, making AI toolkits easier for the community to use. This role offers an outstanding opportunity for someone with an AI background to advance the state of the art in this space.
Responsibilities
- Integrate new communication library features into AI frameworks, from proof-of-concept to production.
- Analyze AI workloads and frameworks to identify multi-GPU communication needs and opportunities.
- Collaborate with teams working on the latest AI models.
- Improve AI compilers to optimize communication or perform automatic fusion.
- Characterize the performance of AI workloads on multi-GPU clusters.
- Design fault-tolerant and elastic solutions for large-scale or dynamic AI workloads.
- Author custom communication or fused compute-communication kernels for optimal performance on NVIDIA platforms.
- Influence the roadmap of communication libraries like NCCL and NVSHMEM.
- Collaborate with a dynamic, multi-time zone team.
Requirements
- B.S., M.S., or Ph.D. in Computer Science or a related field (or equivalent experience).
- 8+ years of software engineering and HPC/AI experience.
- Experience developing or integrating with Deep Learning Frameworks like PyTorch and JAX.
- Experience with Inference Engines such as TRT-LLM, vLLM, and SGLang.
- Rapid prototyping and development experience with Python, C++, CUDA, or related DSLs (Triton, cuTe).
- Solid understanding of AI models, parallelisms, and/or compiler technologies (e.g., torch.compile).
- Experience conducting performance benchmarking on AI clusters.
- Familiarity with at least one performance profiler toolchain (e.g., PyTorch profiler, NVIDIA Nsight Systems).
- Understanding of HPC/AI communication concepts (e.g., 1-sided vs. 2-sided communication, elasticity, resiliency, topology discovery).
- Adaptability and passion for learning new areas and tools.
- Flexibility to work and communicate effectively across different teams and time zones.
Benefits
- Equity
- Benefits