CareerPlanSign in

Principal Software Engineer

United States, Washington, Redmond💼 Full-time🗓 2026-08-04 → 2026-09-26

Core

Design and build GPU-accelerated infrastructure for training and inference workloads, managing device scheduling, isolation, and sharing across bare metal, VMs, and containers.

Role type

Principal Software Engineer (GPU Infrastructure & Orchestration)

Builds

GPU device management systems, orchestration platforms (AKS, DRA, Kubernetes), virtualization stacks, and high-performance interconnects for AI workloads.

Domain

Cloud Infrastructure, AI/ML Platforms, GPU Computing, Virtualization

Deliverable

production ML models | infrastructure

Required skills

GPU device management, Kubernetes ecosystem, high-performance networking (RDMA/InfiniBand), virtualization and containerization, system design for large-scale fleets, observability and diagnostics, technical leadership, cross-team architectural alignment, debugging complex cross-layer systems.

Preferred skills

Experience with confidential compute, multitenant AI platforms, GPU virtualization/passthrough/partitioning technologies.

Technologies

AKS, Dynamic Resource Allocation (DRA), Kubernetes, C, C++, C#, Java, JavaScript, Python, RDMA, InfiniBand

Responsibilities

Design and build GPU-accelerated infrastructure for training and inference; develop systems for GPU device management, scheduling, isolation, and sharing; build and operate advanced orchestration and resource governance scenarios; evolve virtualization and container stacks for modern AI workloads; optimize performance and utilization across large GPU fleets; partner with networking and storage teams for high-performance interconnects; drive end-to-end platform features including observability and operational excellence; influence platform architecture and technical direction through design reviews.

Seniority

Principal, hands-on IC with strategic influence and mentorship

Sourced via microsoft · Listed on CareerPlan, which tracks 70,000+ jobs from 20+ sources.