This role pays less than
55%
of similar roles. Most pay
$162,000–$231,362
— the shaded band above.
At the midpoint, this role pays about
$184k
versus about
$197k
for comparable roles.
Based on 240 similar postings.
Employer
About Broadcom
Broadcom is a global semiconductor and infrastructure software company that designs and markets a wide range of networking, storage, and wireless connectivity solutions. Industry: Semiconductors & Infrastructure Software
Broadcom currently has
119 open roles
on FindRole.
Listed pay typically runs
$121,900–$195,000
across 105 roles with salary data.
As an AI Systems Performance Engineer within the Performance Lab, you will drive performance benchmarking for AI inference, training, and storage workloads. You will focus on optimizing the Ethernet fabric connecting servers to ensure seamless data flow for distributed AI workloads. Your daily responsibilities include executing rigorous benchmarks like MLPerf and NCCL tests, isolating complex system bottlenecks across Linux OS, server hardware, and Ethernet switches, and developing automated testing frameworks. To succeed, you must possess deep expertise in Linux systems, Python, C++, and PyTorch, while demonstrating a strong understanding of how AI models consume compute and network resources. You will also utilize RDMA, RoCEv2, Docker, and Kubernetes to troubleshoot distributed systems. This role solves critical performance puzzles by tuning network parameters to achieve maximum throughput and minimum latency for large-scale machine learning infrastructure.
Python
C++
Go
Rust
Kubernetes
Ray
Slurm
vLLM
SGLang
Triton
PyTorch
JAX
ROCm
HIP
CUDA
MLflow
Weights & Biases
CI/CD
Distributed Systems
GPU Infrastructure