Staff GPU Inference QA & Reliability Lead

Cerebras; United States; Remote; USD 180,000 - 250,000; Full time

Job summary

Cerebras Systems is hiring a Staff GPU Inference SDET to lead the validation and reliability of our GPU inference stack and rack-scale systems. You will design automated test ecosystems for multi-node clusters, ensure numerical correctness, and drive production-grade reliability for high-speed inference workloads.

You will architect and implement validation pipelines spanning API services, serving workers, runtimes, and firmware, while measuring TTFT, ITL, throughput, and tail latency to guide

Responsibilities

• Build GPU Release Qualification Systems: automated test frameworks, regression gates, and release pipelines for the GPU inference stack.
• Inference Serving Workload Validation: benchmark and stress-test distributed LLM serving frameworks, focusing on prefill vs decode, batching, caching, and parallelism.
• Performance Modeling Verification: automate workload replay/benchmarking to validate GPU performance models; track TTFT, ITL, throughput, and P99 latency.
• Quality Gates: ensure model accuracy, precision stability (FP16/FP8/quantization), determinism, and output correctness across updates.
• Fleet Resilience: chaos engineering and fault-injection for multi-node GPU clusters.
• Observability: CI/CD Integration fidelity across the stack; monitor, alert, and report health

Job description

Cerebras Systems is hiring a Staff GPU Inference SDET to lead the validation and reliability of our GPU inference stack and rack-scale systems. You will design automated test ecosystems for multi-node clusters, ensure numerical correctness, and drive production-grade reliability for high-speed inference workloads.

You will architect and implement validation pipelines spanning API services, serving workers, runtimes, and firmware, while measuring TTFT, ITL, throughput, and tail latency to guide