Free the experts.
Automate the rest.
Hi, I’m Liam. I build software that gives skilled people more time to do the work only they can do.
Australian software engineer and architect, based in Utrecht, the Netherlands — building production software since 2015.
The problem I kept seeing
Across the teams I’ve worked with, the same pattern keeps turning up. The person who understands a process best ends up holding it together — answering repeated questions, copying figures between systems, chasing approvals, checking routine work, and preserving fragile workflows through memory and goodwill.
It’s important work, and it falls to them because they can be trusted to get it right. But it isn’t the work that needs their judgement. It’s the work getting in their way.
It is not the person who needs replacing. It is the repetitive work surrounding their judgement.
Automation should support judgement, not replace it
A lot of automation gets sold as a way to remove people from a process entirely. In practice, that’s rarely what teams want. They want the repetitive parts handled so the important decisions stay with someone who understands them — with the work still visible, still reviewable, and still easy to escalate when something unusual happens.
A principle I work by
The better someone is at their job, the less time they should spend on work that does not need their judgement.
That’s how I approach it. I help teams find the work that’s genuinely worth automating and build systems they can understand, operate and improve — not black boxes they have to trust on faith.
What that looks like in practice
The technology changes; the goal stays the same: take avoidable friction off people’s plates. Depending on where that friction lives, that might mean:
- Fragile workflows
- Replace processes held together by browser tabs, spreadsheets and institutional memory.
- Internal tools
- Make routine work boring and repeatable so it stops demanding expert attention.
- AI-assisted workflows
- Build drafting, sorting and summarisation around explicit review and escalation.
- System connections
- Join tools that were never designed to communicate.
- End-to-end delivery
- Take an unclear problem through architecture and into production.
Built in the real world
Since 2015, I’ve built production software across internal platforms, developer infrastructure, policy governance and product interfaces.
At ING, I work on systems inside a large regulated organisation, where reliability, explainability and maintainability matter as much as the initial idea.
I also build open-source tools in Rust, including raff and signal-kit; I wrote tree-sitter-wit, a WIT parser I later donated to the Bytecode Alliance.
When I’m not building
I’m Australian, I live in Utrecht with my fiancée, I play squash, and I’m usually building another side project.
The same instinct runs through all of it: I’d rather quietly remove a small daily annoyance than learn to live with it.
Tell me what’s wasting your team’s time
Have a workflow that leans too heavily on copying, chasing, remembering, or explaining the same thing repeatedly? Send me a short description. I’ll tell you honestly whether automation would help — and if it wouldn’t, I’ll say that too.