1 per engagement
Platform & MLOps
Platform & MLOps Engineer
Seniority band: 8–14 years1 per engagement
Keeps inference fast and cheap: GPU scheduling, model serving, rollout safety and the observability that tells you at 2am which agent misbehaved.
What this role owns
- Serving infrastructure and GPU efficiency
- Progressive rollout and rollback for prompts and models
- Tracing across an agent run, so a failure has a cause not a guess
- Infrastructure cost and capacity planning
Tooling and capability
Agentic AI
Model servingvLLMQuantisationAgent tracing
Software
PythonGoTerraformHelm
Platform
KubernetesGPU schedulingGrafanaPrometheusAWS
