Skip to content
Back to portfolio

Cloud Infrastructure

Cloud Health Alerting

An organization-wide cloud health alerting platform that routes AWS Health events to the right teams before they become incidents.

Senior AI Engineer at Eli Lilly and Company · January 2026 – Present

Public descriptions use a simplified conceptual architecture and omit internal account structures, ownership data, endpoints, and operational details.

Problem

AWS Health events needed to reach the correct internal customer teams, account owners, or service owners without overwhelming them with duplicate or low-value notifications.

My contribution

Worked on SLO-driven routing, event ingestion via EventBridge, normalization, severity-aware filtering, deduplication, ownership-based routing, incident triage automation, and notification-channel evaluation across Outlook and Microsoft Teams.

Simplified architecture

Simplified: AWS Health events flow through EventBridge into an ingestion stage, then normalization, deduplication, ownership resolution, and severity classification, before SLO-driven routing delivers each alert to the owning customer team over Outlook or Microsoft Teams.

Outcomes

  • Targeted at least 95 percent production coverage. (target, not a measured result)
  • Targeted critical alert delivery in under three minutes. (target, not a measured result)

Technologies

AWS HealthAmazon EventBridgeOutlookMicrosoft Teams