Production platform and migrations
Alongside the two systems, I owned the team's production services and pipelines across dev, test and prod, and the engineering practice around them.
- Role
- Production owner
- Context
- Platform engineering
- Stack
- GCPAirflowKubernetesBigQueryGitHub ActionsSecurity scanningSecrets management
- ownership of the team's production services and pipelines
- Dev to prod
- orchestration, CI/CD and cloud environment, with production running throughout
- Migrated
- releases, behind security scans, coverage checks and managed secrets
- Gated
- in an engineering handbook, verified against the live systems
- Documented
This is a public summary. Figures and internal specifics from CVS Health/Aetna are left out; I am glad to go into more depth in conversation.
- OrchestrationAirflow 2toAirflow 3A major-version upgrade of the orchestrator, with every pipeline moved across.
- CI/CDJenkinstoGitHub ActionsA change of CI/CD platform: builds and deployments moved to shared, reusable workflows.
- CloudOld environmenttoNew environmentCompute moved to a new cloud environment on GCP.
An API that is safe to call twice
I rebuilt a data-delivery API so that it is safe to call again and again: scoped access, rate limits and idempotent delivery. Streaming the response ended the timeouts on its largest feed.
Security and quality gates
Hardened the release path: security scans and test-coverage gates that block a release, and secrets held in a managed vault.
Data promotion across environments
Built promotion pipelines that move production tables between environments for other teams, including a multi-region copy.
Documentation and handover
Wrote the engineering handbook the team uses for onboarding and support, and verified it against the live systems. Served as the bridge to a partner organization.