Energage (formerly Engagedly)
DevOps Engineer
January 2026 – Present
Energage (formerly Engagedly)
- Own the Kubernetes platform across multiple EKS clusters and AWS accounts at 99.9% uptime, with live cluster upgrades and Karpenter autoscaling for zero-downtime releases, at around 30% lower infrastructure cost.
- Built the shared CI library nearly every service deploys through, cutting per-service pipeline code by around 70%. Verified the migration by machine, catching real divergence in roughly 46% of pipelines manual review had passed.
- Moved a large microservice fleet onto GitOps without a single redeploy, seeding desired state from live release values, and made per-service rollback a one line config change.
- Run a self-hosted observability platform for logs, metrics and traces in every environment. Audited dashboards by executing their queries rather than reading them, rebuilt the silently broken ones, and cut metrics nobody visualised by around 85%.
- Root-caused a silent recurring outage where ingestion stopped while every health check stayed green, traced it to inode exhaustion rather than disk space, and built the alerting that would have caught it.
- Upgraded and migrated stateful production infrastructure with zero data loss and under a minute of user-visible downtime, every phase gated on a rendered-manifest diff and a backup.
- Replaced long-lived static keys with keyless per-application CI identities after sweeping builds, registries and scripts for embedded credentials, and put human access to cloud APIs behind VPN-only egress.
- Brought hand-built cloud accounts under infrastructure as code without mutating a live resource, each wave gated on a plan proving zero change, after catching tooling that would have replaced a running production database.
- Built a reusable SSO gateway as an OIDC reverse proxy that put internal tools behind corporate identity without code changes, and led access governance across AWS accounts.
- Ran a cost-optimisation programme across the AWS estate cutting NAT data-processing by 83% at no infrastructure cost, and reclaiming spend from backup retention and unattached storage.
- Swept the application fleet for secrets and personal data leaking into logs classified what each service exposed by class of leak, and drove the remediation.
- Root-caused a production latency and error-rate incident end to end confirming each hypothesis against evidence before moving on, then wrote the trail up as a reusable read-only investigation method the team follows.
-GRmdL18E.webp)
