The role involves building and operating production systems to manage and optimize a large-scale accelerator fleet. Responsibilities include developing data pipelines to ingest and normalize telemetry from heterogeneous cloud environments. The engineer will create observability tooling to provide real-time visibility into fleet health and performance. Additionally, the position requires instrumenting workloads to measure hardware efficiency and building internal platforms for capacity planning and cost attribution. The work necessitates collaboration across research, infrastructure, and finance teams to ensure effective resource allocation.

Candidates must have at least five years of software engineering experience with a focus on operating production systems. Proficiency in Python, SQL, and Kubernetes at an operational scale is required, along with experience in data pipeline engineering and observability tooling. A bachelor’s degree or an equivalent combination of education, training, and experience is the minimum educational requirement. Strong communication skills and the ability to navigate high-autonomy environments are essential for cross-team collaboration.