You'd think outgrowing your job means you stop doing it

You’d think outgrowing your job means you stop doing it. For me it meant the opposite.

I still pick up sprint tickets, ship customer-facing code and hit product deadlines. What changed: that work no longer fills the week. The AI agent tooling I built gave me back enough time to take on multiple sets of responsibilities alongside the first.

Now, in parallel with sprint work, I own:

• An end-to-end Playwright regression suite: 263 tests across 41 specs, run every weekday against prod and staging
• Shipping on web, iOS and Android: three codebases, three languages, three build systems
• 100+ reusable agent skills in five bundles that my 12-person team installs, and now contributes to
• A multi-agent plugin I published to our internal plugin marketplace, so anyone can install it

The surprise

Once output volume goes up, its nature changes.

Shipping more of your own tickets is throughput. Spend freed time on things other people use (a shared test suite, a release playbook, an installable plugin) and the work stops being about how much you ship. It becomes about how much everyone around you ships.

That’s when volume turns into influence.

As a contractor I don’t have a promotion ladder, so I borrowed the one my full-time peers use. It draws lines by scope, not volume: Senior delivers within a team; Staff owns something across teams and sets standards. More PRs barely move you. Adoption does.

What I’d tell another engineer

• Don’t give freed time back to the backlog by default.
• Keep the day job airtight; credibility funds everything else.
• Measure adoption, not just volume.
• Package it: work that only runs on your machine is a hobby, work anyone can install is leverage.

AI tooling is sold as a way to do the same job faster. The more interesting outcome is doing the same job and a bigger one.

Next year’s goal isn’t more PRs. It’s more people shipping faster because of something I built.