DevOps & Site Reliability Engineer: I design infrastructure that stays calm under load.
Environments online: DEV, QA, SB, STG, PROD, PROD2
Most engineers pick a side: stable or creative. I work on both planes at once. Scroll and watch the work travel from pure stability towards the top-right corner, where it becomes OG.
Every deploy changelog is parsed, matched against what really exists in AWS, and scored before a human ever reviews it.
Self-built ECS health and cost dashboard: every service across every environment, overloaded ones flagged first, memory-leak hints, 6-hour trends. The CEO praised it publicly.
Heartbeat monitor: steady pulse, about 72 beats per minute, across 6 services.
100+ containers moved off memory-starved 8 GB EC2 boxes (constant out-of-memory crashes) onto right-sized ECS with autoscaling, inside the old budget. Then the whole fleet went x86 → arm64.
One-Click Deployment, which I led with a team of 3. An orchestrator reads a formatted changelog, decides create-or-update for every ECS, Lambda, API and DynamoDB change, and dispatches it. It replaced releases that needed 5–6 engineers by hand.
Including the ECS module the whole team still builds on. New environments went from hand-built to a few inputs, and the 6+ environments stopped drifting apart.
100+ containers off crash-prone 8 GB EC2 machines, one task per service, CPU and memory autoscaling, all inside the old EC2 budget.
100+ services × 6 environments: Dockerfiles, base images and build pipelines rebuilt for arm64.
Parent-child dynamic GitLab CI/CD: deploy any service to any of 6 environments from a single place.
The ECS health and cost dashboard the team watches. The CEO shared it publicly on LinkedIn.
A team of 3; changelog-driven automation that replaced error-prone releases run by 5–6 engineers.
Still curious, still shipping originals.
✓ HASHICORP TERRAFORM ASSOCIATEAWS SOLUTIONS ARCHITECT · IN PROGRESS*A DevOps engineer built this. Frontend folks, no pressure.