Senior GPU Inference Engineer for Real-Time AI

Cerebras; United States; On-site; USD 150,000 - 210,000; Full time

Job summary

Cerebras Systems is seeking a Software Engineer to productionize and optimize our GPU serving stack. You will work across custom inference APIs, vLLM, ROCm, PyTorch, and AMD GPU infrastructure to ensure reliable, numerically correct, observable, and high-performance serving.

You will write production code, establish operational practices for a new accelerator fleet, and push improvements in time to first token, throughput, tail latency, and capacity efficiency.

Qualifications

• 5+ years of software engineering experience with ownership of complex production systems.
• Experience building or optimizing production inference systems for large GPU workloads.
• Strong programming in C++ and Python, with multithreading and memory management.

Responsibilities

• Productionize the GPU inference stack across APIs, vLLM, PyTorch, ROCm, and GPU infrastructure.
• Own GPU operational readiness including deployment, upgrades, health checks, and rollback strategies.
• Drive reliability with SLIs/SLAs, automated recovery, incident response, and post-incident remediation.

Skills

C++

Python

Distributed systems

Performance optimization

GPU inference

Education

Bachelor's degree in CS/CE/EE

Tools

vLLM

PyTorch

ROCm

TensorRT-LLM

Kubernetes

Job description

Cerebras Systems is seeking a Software Engineer to productionize and optimize our GPU serving stack. You will work across custom inference APIs, vLLM, ROCm, PyTorch, and AMD GPU infrastructure to ensure reliable, numerically correct, observable, and high-performance serving.

You will write production code, establish operational practices for a new accelerator fleet, and push improvements in time to first token, throughput, tail latency, and capacity efficiency.