Own production reliability for AI/ML services by monitoring availability and performance, defining SLOs and SLIs (latency, error rates, model performance) tied to user experience and business impact, and owning recurring model refresh cycles (retraining, validation, and redeployment) to prevent staleness and drift. Provision, deploy, configure, and maintain new AI infrastructure as it is adopted, including AI gateways, RAG platforms, LLM observability tools, agentic workflows, and no-code agent builders, using tools such as Union Cloud, BentoML, Feast, Kubernetes, Terraform, and GCP services.