Fifteen years underneath the product.
I’m a senior platform engineer in Sydney. I work on the layer most people never see. It’s the part that decides how safe a change feels before anyone ships it.
- Based
- Sydney, Australia
- Role
- Senior platform & DevSecOps engineer
- Stack
- GCP, AWS, Kubernetes, Pulumi, GitHub, Cloudflare
- Open to
- Contract and consulting work
What I do
Most of it is boring, by design: keeping access tight, getting changes reviewed before they ship, and making sure problems show up early. When that’s in place, changing production stops being a big deal.
I started on the contractor side, the engineer a consultancy sends in to fix other companies’ infrastructure, one after another. Then in-house and close to the hardware at a national bank: virtualisation, storage, backups. An outage there got counted in the number of people who couldn’t do their jobs. From there into reliability engineering, and now senior platform and DevSecOps roles at startups, where the problem inverts: move fast, but leave something auditable behind instead of a pile of one-off hacks.
GCP is home base, with AWS and on-prem when the work calls for it. The tools turn over every few years; what carries over is knowing where complexity is worth paying for.
The most recent of those was a rebuild: a live cloud platform whose ownership had collapsed into its folder layout, re-cut into tiers that each change for their own reason, with typed contracts between them and release checks that stop a bad change before it applies. I wrote that one up in full as a series, Making IaC boring.
How I work
Simple by default: managed services and proven primitives before anything bespoke, and a custom tool only when the problem genuinely needs one. Most of what I walk into that looks complex is an old shortcut that earned a nicer name.
An awkward conversation about a risk is cheaper before a release than during an incident, so I have it early. I’d rather write the dull runbook than be the only person who knows how something works.
I’ve led teams and owned environments end to end. The handover is what I judge the work by: whether the next engineer can read what they inherited and change it without fear.