Tech Stack

Description

You will own the reliability, performance, and observability of the entire inference stack at Nebius Cloud, designing telemetry pipelines, tuning Kubernetes autoscalers, and hardening request-routing to ensure flawless behavior under extreme load. Your work directly scales a platform that deploys text, vision, audio, and multimodal models at massive scale.

Requirements

• Deep fluency with Kubernetes, Prometheus, Grafana, and Terraform
• Proficiency in Python or Bash scripting
• Experience with alert design and SLOs for high-throughput APIs
• Experience with GPU-heavy workloads (vLLM, Triton, Ray, or similar)
• Background in MLOps or model-hosting platforms

Responsibilities

• Design and refine telemetry pipelines for metrics, logs, and traces
• Tune Kubernetes autoscalers for GPU efficiency
• Craft Terraform modules for resilient cluster deployments
• Harden request-routing and retry logic for transparent failure handling
• Drive incident detection, isolation, remediation, and post-mortem culture