You’ll scale our CUDA / Triton inference fleet across regions and providers, keeping costs low and QoS high.

Responsibilities

• Design autoscaling policies for GPU and CPU inference clusters.
• Implement canary & shadow traffic pipelines for new model versions.
• Drive infra-as-code and chaos testing culture across engineering.

Minimum Qualifications

• 4+ years Go, Rust or similar in production micro-services.
• Strong grasp of container orchestration (K8s, Nomad) and service discovery.
• Experience with GPU scheduling or network-heavy realtime systems a plus.
• BS in CS, CE or demonstrable equivalent.