Mainpac
Full-timeCurrentSoftware Engineer
to Present (11 mo)
- Took the team's AI workflow maturity from Level 1 to Level 2 with project-specific Claude skill sets, and set the Level 3 target of agents that triage, analyse, test, fix, and monitor.
- Cut environment bake from 45 minutes on a developer machine to 25 on a Kubernetes cluster, and made those bakes parallel, so ten fix branches now deploy and test at once in less wall-clock time than one branch took locally.
- Built a workflow that submits test jobs and returns results automatically, so any branch gets a real deployed environment rather than a queue slot, and an agent gets the feedback loop it needs to judge its own change.
- Set up agent workflows that author unit and end-to-end tests around a reported defect, so the suite reproduces the bug before the fix lands and proves it afterwards.
- Automated new EAM screen and grid development, the most frequent category of client requests, with a Claude skill set that directs implementation end to end.
- Standardised end-to-end test authoring with a four-skill Claude workflow: test design, edge-case and equivalence-class exploration, implementation, and step verification against Mainpac user guides.
.NET 8/10
ASP.NET Core
.NET MAUI
Claude skill authoring
Kubernetes
- Agentic workflow design
- AI-driven test generation
