Lynes

01

Ground zero is a strategy problem before it is a tooling problem

The first job was deciding what the transformation was going to look like across several legacy teams with different habits, then choosing a toolchain that fit them (GitHub Actions and Codefresh), and replacing the manual build and deployment processes team by team.

02

Reusable modules, so the second service is cheap

I established the Infrastructure as Code architecture on Terraform and Terragrunt, and the reusable modules that came out of it became the foundation every service was built on. On top of that went the Kubernetes migration onto AWS EKS, with GitOps workflows that let developers deploy without going through me.

03

Self-service only works if someone can see what broke

Observability on Prometheus, Grafana and CloudWatch, with PagerDuty wired in for on-call management. Alongside it, mentoring teams in cloud-native practice, the part that decides whether any of this outlives the person who built it. Two and a half years in-house is long enough to find out; the patterns are still in use.

Engagement

2.5 years

Started from

Manual deploys