Senior GPU Inference Systems Engineer

Cerebras; Sunnyvale (CA); On-site; USD 180,000 - 280,000; Full time

Job summary

Cerebras Systems is hiring a Software Engineer to productionize and optimize the GPU serving stack across custom inference APIs, vLLM, and ROCm-based infrastructure. This hands-on role emphasizes performance, observability, and reliability in a disaggregated AI inference environment.

You will work across application, runtime, distributed systems, and hardware layers to boost time-to-first-token, throughput, tail latency, and capacity efficiency on the Cerebras Wafer-Scale Engine.

Qualifications

• Minimum 8+ years of software engineering experience in complex production systems.
• Experience building/operating/optimizing production inference systems for large language or GPU workloads.
• Strong programming in C++ and Python with attention to concurrency and performance.

Responsibilities

• Produce and optimize the GPU inference stack from API services to model-serving workers.
• Ensure GPU fleet operational readiness with deployment, upgrades and rollback strategies.
• Define SLIs/SLAs and improve fault isolation, automated recovery, and incident response.

Skills

C++

Python

Multithreading

Distributed systems

GPU performance

Linux

Kubernetes

Benchmarks

Communication

Leadership

Education

Bachelor’s degree in CS/CE/EE

Tools

vLLM

TensorRT-LLM

Triton Inference Server

HIP/ROCm

Docker

Job description

Cerebras Systems is hiring a Software Engineer to productionize and optimize the GPU serving stack across custom inference APIs, vLLM, and ROCm-based infrastructure. This hands-on role emphasizes performance, observability, and reliability in a disaggregated AI inference environment.

You will work across application, runtime, distributed systems, and hardware layers to boost time-to-first-token, throughput, tail latency, and capacity efficiency on the Cerebras Wafer-Scale Engine.