Senior Deep Learning Communications Architect — Equity Options

NVIDIA; Redmond (WA); On-site; USD 184,000 - 357,000; Full time

Benefits offered by this job

Equity

Benefits package

Job summary

NVIDIA is seeking a Deep Learning Communication Architect in Redmond to scale DNN models and DL training/inference frameworks across large clusters. You will optimize data transfer, design protocols, and collaborate on hardware interconnects such as NVLink/InfiniBand to reduce latency and improve scalability.

The role requires advanced degrees and substantial experience with DL frameworks and parallelism techniques. Competitive salary, equity, and comprehensive benefits are offered.

Qualifications

• Ph.D./Masters/BS in CS, EE, CSEE or related field or equivalent experience.

Responsibilities

• Scale DNN models and training/inference frameworks to systems with hundreds of thousands of nodes.
• Optimize communication performance and reduce bottlenecks during distributed DL training/inference.
• Design efficient communication protocols for DL workloads to minimize overhead and latency.
• Collaborate with hardware/software teams to apply high-speed interconnects and libraries like MPI, NCCL, UCX, NVSHMEM.
• Explore new communication technologies to improve performance and scalability of DL systems.
• Build proofs-of-concept, run experiments, and model quantitatively to deploy new strategies.

Skills

DNN development

LLM performance optimization

Data parallelism

Pipeline parallelism

Tensor parallelism

Expert parallelism

FSDP

Disaggregated serving

Dynamo/Triton

PyTorch

TensorRT-LLM

vLLM

SGLang

C++

Python

CUDA

OpenCL

InfiniBand

RoCE

Education

PhD, Masters, or BS in CS/EE/CSEE or related field

Tools

PyTorch

TensorRT-LLM

vLLM

SGLang

CUDA

OpenCL

Job description

NVIDIA is seeking a Deep Learning Communication Architect in Redmond to scale DNN models and DL training/inference frameworks across large clusters. You will optimize data transfer, design protocols, and collaborate on hardware interconnects such as NVLink/InfiniBand to reduce latency and improve scalability.

The role requires advanced degrees and substantial experience with DL frameworks and parallelism techniques. Competitive salary, equity, and comprehensive benefits are offered.