The opportunity
Cerebras Systems builds the world's largest AI chip, 56 times larger than GPUs. This architecture allows Cerebras to deliver industry-leading training and inference speeds; over 10 times faster than GPU-based hyperscale cloud inference services.
What you'll do
Design and optimize scalable data pipeline architectures for multi-terabyte: hardware telemetry, reliability analytics, and performance optimization.
Architect, develop, and optimize hyperscale data pipeline frameworks and ETL: processes to aggregate, process, and analyze multi-terabyte hardware performance and telemetry streams, including utilization, power, thermal, acoustic, and reliability metrics across heterogeneous compute, storage, and AI server platforms, ensuring hardware performance compliance and operational reliability.
Design and implement hardware performance analysis and anomaly detection: systems using Python, SQL, Tableau, Hive, and Spark to forecast hardware failure curves, identify performance bottlenecks, and generate prescriptive recommendations for hardware and system optimization.
Lead hardware characterization experiments and thermal/cooling A/B studies to: evaluate operational envelopes, delivering validated strategies that reduce carbon footprint, improve water usage efficiency, and maintain or enhance system reliability.
Engineer telemetry ingestion, monitoring, and visualization systems to: provide real-time, high-fidelity hardware health data to hardware, firmware, and datacenter operations teams, enabling data-driven decision-making at scale.
Define, operationalize, and maintain custom efficiency and reliability: metrics; perform root cause analysis of systemic failures using large-scale statistical and machine learning methods; and deploy solutions that improve platform scalability, energy efficiency, and sustainability.
What they're looking for
- Collaborate with cross-functional engineering teams to troubleshoot complex: failures, isolate defective components, and implement systemic fixes across CPU, GPU, DRAM, PCIe, networking, and storage subsystems.
- Support the evolution and optimization of next-generation AI platforms and: silicon products, including hardware subsystems (CPU, GPU, DRAM, PCIe, networking, and storage), to meet the performance, scalability, and efficiency demands of large language model training and inference workloads.
- Large-scale data pipeline architecture and ETL, distributed data processing: (Hive, Spark), and dashboard development;
- Python, SQL, Tableau, Linux, and automation scripting;
