Software Development Engineer — GPU Fleet Management & AI Infrastructure
Core
Design and build Fleet Manager, a secure control plane for scheduling, workload orchestration, and model-serving across large-scale AMD GPU clusters.
Role type
Senior IC systems software engineer (GPU infrastructure)
Builds
Distributed control planes, Kubernetes schedulers, inference gateways, and developer-facing APIs for GPU training and inference
Domain
Cloud infrastructure, accelerated computing, distributed systems
Deliverable
production ML models
Required skills
Rust, C++, Go, distributed systems, Kubernetes internals, PostgreSQL, concurrency, asynchronous programming, failure recovery, REST/gRPC APIs, observability, security, multi-tenant isolation
Preferred skills
ROCm, HIP, GPU topology, vLLM, SGLang, PyTorch, container security, threat modeling, bare-metal/virtualized GPU infrastructure
Technologies
Kubernetes, Kueue, JobSet, Slurm, Spur, PostgreSQL, ROCm, gRPC, WebSocket, Prometheus, Grafana
Responsibilities
Design distributed control-plane services and schedulers; build orchestration for GPU training and inference; implement admission control and topology-aware placement; develop durable reconciliation and lifecycle management; integrate with container runtimes and observability platforms; build secure multi-tenant infrastructure; diagnose complex distributed failures; lead technical projects and mentor teams
Seniority
Senior, hands-on IC with technical leadership
