Principal Engineer Capacity/Compute
Core
Build planning and allocation systems for cross-region and cross-provider placement, guardrails, queueing, and occupancy KPIs; drive compute efficiency programs covering stranding, rightsizing, unused capacity recovery, benchmarking, and workload utilization across training, inference, and evaluation systems.
Role type
Principal Engineer, Capacity & Compute Infrastructure
Builds
Production systems for compute planning, allocation, efficiency optimization, and cost attribution across multi-cloud environments.
Domain
Cloud Infrastructure / Compute Operations / Data Engineering
Deliverable
production ML models | infrastructure
Required skills
Python, SQL (BigQuery), Kubernetes operations, Cloud provider operations (AWS/GCP/Azure), Prometheus/PromQL/Grafana, Data pipeline design, SLO ownership, Requirements gathering
Preferred skills
Capacity planning, Resource management, Cost attribution, Scheduling and packing efficiency, Large-scale ML infrastructure, Multi-cloud data ingestion, Total cost of ownership (TCO) forecasting, Accelerator infrastructure, Exabyte-scale storage efficiency
Technologies
BigQuery, Kubernetes, Prometheus, Grafana, AWS, Google Cloud, Microsoft Azure
Responsibilities
Build planning and allocation systems for cross-region and cross-provider placement; Drive compute efficiency programs covering stranding, rightsizing, unused capacity recovery, benchmarking, and workload utilization; Own compute cost attribution and forecasting by reconciling provider billing, telemetry, internal systems, demand signals, and research roadmaps; Build and operate data pipelines ingesting occupancy, utilization, and cost data into BigQuery; Operate Kubernetes-native collection, labeling, reservation, taint, and scheduling systems at scale.
Seniority
Principal, hands-on IC
