Perplexity
Member of Technical Staff (Software Engineer, GPU Cluster Infrastructure)
$250K–$485K
San Francisco +4 · Staff · Full-time · Posted Oct 28, 2026
- Location
- San Francisco
Remote (United States)
Toronto
Seattle
New York City
About the job
Own and build a self-serve platform for provisioning and operating GPU infrastructure across cloud providers for AI training and inference workloads.
What you'll do
- Build a platform for launching training jobs and operating inference services without managing underlying GPU infrastructure.
- Operate and optimize the multi-provider GPU fleet, including provisioning, scheduling, reliability, and capacity management.
- Develop Kubernetes orchestration, fault tolerance, autoscaling, and observability for training and inference workloads.
What they ask for
- Deep Kubernetes experience, including custom operators, CRDs, and multi-cluster federation.
- Experience managing GPU clusters at scale, including NVIDIA hardware, CUDA, and InfiniBand or RoCE networking.
- Experience orchestrating compute across multiple clouds and strong distributed-systems fundamentals.
- Infrastructure or systems-level coding in Go, Rust, or C++; experience supporting both training jobs and high-availability inference services.
Nice to have
- Inference serving stacks such as vLLM, SGLang, or TensorRT-LLM; Slurm or other HPC schedulers.
- GPU kernel work in CUDA or Triton; production experience with InfiniBand, RoCE, or RDMA; ML observability tools such as Prometheus, Grafana, or Weights & Biases.
Pay
$250K – $485K (USD, base salary)
Tools
Kubernetes, NVIDIA, CUDA, InfiniBand, RoCE, CoreWeave, AWS, GCP
AI focus
Build infrastructure for AI training and inference workloads, including GPU scheduling, orchestration, and serving.
Model-written summary · the original posting is authoritative