- Operate and improve secure, reliable ML training and serving infrastructure on GKE and Vertex AI.
- Optimize GPU usage, capacity, quotas, and costs.
- Build reusable templates, CI/CD pipelines, and self-service tools for independent model deployment by data scientists.
- Manage model registry, versioning, deployment, rollback, drift monitoring, and serving health.
- Develop workflows using Kubeflow or Vertex AI Pipelines.
- Collaborate with data engineers on feature and training data flows across BigQuery and Airflow.
- Manage infrastructure as code through GitOps, using Config Sync or equivalent tools.
- Enforce least-privilege access to personally identifiable information and data protected by policy tags.
- Create practical documentation and define service-level objectives (SLOs) for reliable platform operation and adoption.
PythonKubernetesAirflow+2 more