Infra & DevOps
Boring infrastructure: typed IaC, blue/green, observable.
Infra & DevOps is boring infrastructure: typed iac, blue/green, observable.
Why this work matters
Most stacks are one resignation away from collapse. The person who set up the infra left, the docs are out of date, and nobody wants to touch it. We replace tribal knowledge with typed code and runbooks.
The work, in detail.
- Pulumi / Terraform / OpenTofu
- Kubernetes / serverless / containers
- OpenTelemetry across stack
- Incident management + on-call
- Secrets hygiene + rotation
- Disaster recovery + restore drills
- →Typed IaC for full stack
- →Observability dashboards (logs/metrics/traces)
- →Incident runbooks
- →Quarterly DR drill report
Sleep on weekends. Typed infrastructure-as-code, gradual rollouts, blast-radius limits, and observability you can actually navigate at 3 AM.
The approach.
Typed IaC
Every resource lives in code, in a repo, with PRs. Pulumi/Terraform with type checking; reviewable plans before any change touches prod.
Observability before optimization
Logs, metrics, and traces wired into a single pane. We don't tune perf without data — and you don't either, after we hand over.
Restore drills
Backups that haven't been restored aren't backups. We schedule restore drills quarterly and document the runbook live.
Infra & DevOps — common questions
What does typed infrastructure-as-code mean?
Every resource lives in code in a repo, behind PRs, using Pulumi, Terraform, or OpenTofu with type checking. Changes produce reviewable plans before anything touches production, which replaces the fragile setup that lives only in one person's head.
What platforms and tooling do you support?
Kubernetes, serverless, and containers for compute, OpenTelemetry for observability across the stack, plus secrets hygiene with rotation and incident management with on-call. The aim is a stack you can navigate at 3 AM.
How do you approach performance tuning?
Observability before optimization. We wire logs, metrics, and traces into a single pane and do not tune performance without data — and after handover, neither should you have to guess.
Are backups actually tested?
Yes. Backups that have not been restored are not backups, so we schedule restore drills quarterly and document the runbook live. You receive a quarterly DR drill report as a deliverable.
Who is this for?
Teams whose stack is one resignation away from collapse — where the person who set up the infra left, the docs are stale, and nobody wants to touch it. We turn that into typed code, dashboards, and incident runbooks anyone on the team can follow.
More from Software Development
The cost of waiting
is your competitor.
Every 90 days you delay is 90 days of authority compounding for someone else. Get the audit. See the math. Then decide.