Must have
- 4+ years professional WordPress development including custom themes, ACF, and multilingual setups (Polylang/WPML)
- Deep hands-on Pantheon experience: dev/test/live workflow, multidev, Git mode vs SFTP mode, Autopilot, terminus CLI, and the platform's quirks (e.g., update-core hidden on test/live)
- WP Engine deployment and backup workflows
- Strong Git discipline: feature branches, clean merges, understanding why a branch 14 commits behind master is a red flag
- Front-end build pipelines: SCSS, gulp, npm — and the judgment to never commit compiled output without its source
- A documented QA habit: you can describe, step by step, what you check after a plugin update and why
- Clear written English; comfortable communicating directly with agency partners and, when needed, their clients
- Demonstrated fluency working alongside AI assistants (Claude, Copilot, or similar) in a real production workflow — and clear judgment about where AI output must be human-verified
Run monthly plugin, theme, and WP core updates across a large Pantheon/WP Engine roster using multidev and Autopilot branches — never directly on live
Take a manual backup before every live push, without exception, and verify restore points exist before touching production
Perform structured post-deployment QA on live (key templates, archive pages, plugin-dependent modules like Ajax Load More, Polylang, ACF-driven layouts) so breakage is caught by us, not the client
Know and track plugin-specific quirks per site (e.g., plugins that need deactivate/reactivate after core updates, incompatible importers that must be disabled before a core bump) and maintain that knowledge in a living runbook
Work exclusively through source: edit SCSS/JS sources and recompile via gulp — never patch minified files directly; keep branches synced with master and named after their feature
Respect content integrity: never pull databases down over environments carrying client content; never restore backups onto a live or test environment without confirming what will be lost
Deliver monthly maintenance reports on time, in the correct format and folders, with accurate hosting/backup details per site — reports must reflect what actually shipped to live
Triage urgent "site down" flags fast, communicate root cause honestly (including when it was our push), and close tickets only when the fix is verified on live
Use AI tooling (Claude, Claude Code, and similar) as a daily co-worker: drafting QA checklists, diffing changelogs for breaking changes before updates, generating regression test lists per site, summarizing incident post-mortems, and accelerating debugging — while personally verifying everything AI produces before it ships
