General Compute
Platform Engineer
San Francisco +1 · Hybrid · Full-time · Posted Sep 14, 2026
- Location
- San Francisco
NYC
About the job
Builds the inference cloud platform that routes, places, and schedules workloads across mixed GPU and ASIC infrastructure, exposing rack capacity as a reliable, scalable product.
What you'll do
- Build and own the control plane, including fleet routing, model placement, and scheduling across mixed hardware.
- Build the API and serving layer, with reliability and observability, to expose rack capacity as a product.
- Partner with data center deployment and model bring-up teams and contribute to platform architecture.
What they ask for
- Strong systems engineering background in distributed cloud control planes and multi-tenant orchestration at scale.
- Track record building reliability from scratch and comfort as a high-impact individual contributor.
- Experience with resource scheduling or bin-packing for heterogeneous compute pools.
- Experience with multi-cluster federation or distributed consensus systems such as Raft or ZooKeeper.
Nice to have
- Experience with distributed job schedulers or cluster orchestration systems such as Kubernetes, Nomad, or Slurm.
- Experience running non-NVIDIA accelerators in production.
Tools
Kubernetes, Nomad, Slurm, Raft, ZooKeeper, GPUs, ASICs, TPUs
AI focus
Build infrastructure for AI inference, including fleet-wide routing, placement, and serving across heterogeneous hardware; per-model serving performance is owned by the Inference Engineer.
Model-written summary · the original posting is authoritative