Senior Deep Learning Communication Architect

Nvidia

Confirmed live today High trust

Quick summary

Work type: On-site
Location: Santa Clara, CAAustin, TX
Salary: $184,000–$287,500 / yr
Employment: Full-time
Posted: 160 days ago
Freshness: Confirmed live today
Nearby: 99+ roles within 25 mi

Market check

Salary context

Above market

How this pay compares to similar roles

Similar $226k

This role $236k

$172k $300k

below market most similar roles pay here above market

This role pays more than 68% of similar roles. Most pay $196,750–$254,750 — the blue band above. At the midpoint, this role pays about $236k versus about $226k for comparable roles.

Based on 240 similar postings.

Employer

About Nvidia

Nvidia is a leading designer of graphics processing units (GPUs) and system-on-chip units, powering gaming, professional visualization, data centers, and artificial intelligence workloads. Industry: Semiconductors & AI Computing

Nvidia currently has 1623 open roles on FindRole.

Listed pay typically runs $184,000–$287,500 across 1110 roles with salary data.

Most-posted roles

View all roles at Nvidia

At a glance

TL;DR · Senior Deep Learning Communication Architect

Role Posting Log in to save

The Senior Deep Learning Communication Architect joins the software architecture group to scale DNN models and training or inference frameworks across systems with hundreds of thousands of nodes. The role involves identifying bottlenecks in data transfer and synchronization, designing efficient communication protocols, and conducting quantitative modeling to validate new strategies. You will collaborate on hardware and software co-crafting to apply high-speed interconnects like NVLink and InfiniBand while utilizing libraries such as MPI, NCCL, UCX, and NVSHMEM. Required expertise includes parallelism techniques like Data, Pipeline, Tensor, and Expert Parallelism, alongside experience with PyTorch, TensorRT-LLM, vLLM, and SGLang. Candidates must possess strong C++ and Python programming skills, along with proficiency in CUDA, OpenCL, and RoCE networks to optimize state-of-the-art LLM performance.

Skills

Deep Learning PyTorch TensorRT-LLM vLLM SGLang C++ Python CUDA OpenCL NCCL MPI UCX NVLink InfiniBand RoCE Triton FSDP Data Parallelism Pipeline Parallelism Tensor Parallelism Expert Parallelism

What you'll do

Identify and eliminate bottlenecks in data transfer and synchronization during distributed deep learning training and inference.
Develop and implement communication algorithms and protocols tailored for deep learning workloads to minimize latency.
Co-craft systems with hardware and software teams to apply high-speed interconnects and communication libraries.
Research and evaluate new communication technologies to enhance the performance and scalability of deep learning systems.
Build proofs-of-concept and conduct experiments to validate and deploy new communication strategies.
Perform quantitative modeling to optimize the performance of state-of-the-art models on cutting-edge hardware.

What we're looking for

A Ph.D., Masters, or BS in Computer Science, Electrical Engineering, or a closely related field or equivalent experience.
6+ years of experience in building, scaling, or parallelizing DNNs and deep learning training and inference workloads.
Experience evaluating and optimizing LLM training and inference performance on cutting-edge hardware.
Deep understanding of parallelism techniques including Data, Pipeline, Tensor, Expert Parallelism, and FSDP.
Understanding of emerging serving architectures like Disaggregated Serving and inference servers like Dynamo and Triton.
Proficiency in developing code for DNN training and inference frameworks such as PyTorch, TensorRT-LLM, vLLM, or SGLang.
Strong programming skills in C++ and Python.
Familiarity with GPU computing (CUDA and OpenCL) and InfiniBand and RoCE networks.

Save