observability engineer for AI agent platform

The client is a large global enterprise operating across multiple markets, investing in modern cloud, data, and AI capabilities. The project is building an enterprise-grade AI Agent Platform to support the development, deployment, orchestration, security, and observability of AI agents across teams and use cases.

• Design and operate a multi-account New Relic observability environment, consolidating telemetry from multiple project tenants
• Build cross-account NRQL queries, dashboards, and reporting for centralized operational visibility
• Implement dashboards-as-code and alert-policy-as-code using Terraform and Crossplane
• Configure and manage New Relic alert policies and routing to Opsgenie, with future integration into ServiceNow
• Support telemetry ingestion into New Relic through OpenTelemetry/ADOT, OTLP, and Kinesis Firehose pipelines
• Define scalable metric dimensions and low-cardinality facets for large AI agent estates
• Establish observability patterns for Technical, Business, and Audit telemetry segregation
• Support cross-tenant monitoring, reporting, and operational governance
• Collaborate on observability for GenAI workloads, including AWS Bedrock AgentCore telemetry and cost/token attribution
• Automate provisioning and lifecycle management of observability resources using Terraform and Crossplane

• 6+ Years of observability or SRE engineering experience
• Experience operating New Relic or equivalent SaaS observability as a multi-tenant consolidation layer
• Experience with cross-account/cross-tenant telemetry aggregation and reporting
• Experience working within a Technical/Business/Audit observability segregation model
• New Relic administration at scale, including management account and cross-account setup
• NRQL cross-account queries aggregating telemetry from multiple project tenants
• Dashboards-as-code and alert-policy-as-code using the Terraform New Relic provider and Crossplane
• New Relic alert policies routed to Opsgenie
• OpenTelemetry/ADOT fundamentals, including OTLP and Firehose ingestion into New Relic
• Metric dimension design for large agent estates using stable, low-cardinality facets
• Будет плюсом: AWS Bedrock AgentCore observability (Transaction Search, GenAI Observability), Kinesis Firehose metric/log forwarding pipelines (CloudWatch → New Relic), cost/token attribution modeling for GenAI workloads, Crossplane-based provisioning of observability resources

• Benefits supporting well-being and professional growth