Principal Model Optimization Engineer
Core
Design and build GPU and compute management systems for ML teams, covering scheduling, capacity planning, and performance optimization.
Role type
Principal Model Optimization Engineer (ML Infrastructure)
Builds
GPU/compute management systems, training pipelines, workflow orchestration, and developer tools for ML teams.
Domain
Machine Learning Infrastructure / Cloud Computing
Deliverable
infrastructure
Required skills
GPU workload management, Kubernetes scheduling, distributed training optimization, system metrics definition, cross-team technical leadership, SDK/API design, performance debugging, capacity planning.
Preferred skills
Experience with multi-cluster dispatch, autoscaling, mixed precision, quantization, and model serving.
Technologies
Kubernetes, GPU clusters, distributed training frameworks.
Responsibilities
Design platform capabilities adopted by multiple ML teams; Build GPU and compute management systems; Define and improve metrics for training startup time, job success rate, and cost; Investigate and fix slow or failing jobs; Collaborate with ML engineers to improve workflows; Lead design reviews and mentor senior engineers.
Seniority
Principal, hands-on IC with mentorship responsibilities.
