Proof of work
Client names are not always public. The engineering method can be.
Many projects cannot be shown by client name because contracts and NDAs block that. These case studies therefore focus on the useful part: constraints, architecture, tool choices, delivery model, risk handling, and measurable outcome bands.
I remove client identity, screenshots, private URLs, revenue, traffic, and vendor details that could identify the buyer. What stays is the decision logic a technical buyer needs to evaluate my judgement.
A practical breakdown of diagnosing Core Web Vitals issues, third-party scripts, cache boundaries, checkout risk, and post-launch measurement without naming the client.
How I plan a headless migration without exposing a client: content model, URL map, preview workflow, schema continuity, cutover and rollback.
A confidential AI content-ops pattern covering Claude API, prompt governance, human approval, cost ceilings, and data boundaries.
An anonymised migration off TYPO3 onto WordPress with full URL preservation, schema continuity, and editor onboarding.
An anonymised WooCommerce headless build on Cloudflare Pages with the Universal Commerce Protocol as the integration boundary.
An anonymised MCP server in front of WordPress so AI agents draft and translate content under typed schemas, scoped auth, and hard cost ceilings.