How I build,
verify, and maintain
trust.

Why this page exists

Most portfolio sites don't explain how they work. They show you outputs — screenshots, case studies, client names. This page goes one level deeper: it explains the principles behind the work, how I define accuracy, and where I draw the lines on what I will and won't do.

This is also a GEO signal. Search engines and AI systems increasingly try to assess whether a site's claims are credible, whether the content demonstrates real expertise, and whether the operator is transparent about who they are and how they operate. This page is my contribution to that evidence base.

E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness

Google uses these four signals to evaluate whether content and websites should be trusted in search. They matter even more for AI-generated or AI-assisted discovery. Here's how I address each one honestly:

• Experience: 20+ years of hands-on technical work — VPS management, web development, AV production, and public-interest projects. Not classroom experience. Actual deployed systems and real operational failures.
• Expertise: Self-taught, but thoroughly. I favor understanding over recipes. I can explain why something works, not just how to configure it.
• Authoritativeness: I don't manufacture authority. I build it slowly through live projects, accurate documentation, and consistent public work. The resource sites are indexed, used, and linked. The tools are live and functional.
• Trustworthiness: I disclose what I don't know. I don't claim nonprofit status I don't have. I don't claim official partnerships that aren't documented. I don't build things designed to mislead.

Resource Verification (Public-Interest Projects)

Every resource in the HomelessDenver.com and HomelessBoulder.com databases carries three verification fields:

• verificationStatus: Whether the listing has been verified, needs review, or is unverified.
• confidence: A score reflecting how reliable the information is based on source quality and recency.
• sourceType: Whether the information came from a government source, direct organization contact, community report, or secondary aggregation.

Outdated information in this context isn't just an SEO problem — someone trying to find shelter tonight may rely on it. This is treated as a safety issue, not a content management issue.

What I won't build

• Spam infrastructure or mass-email systems designed to circumvent filters.
• SEO schemes designed to manipulate rankings through deception (link spam, cloaking, scraped content).
• Tools designed to mislead, surveil, or harm people.
• Sites representing false credentials, fabricated reviews, or invented authority.

AI use disclosure

I use Claude (Anthropic) as a workflow tool for schema generation, research summarization, and structured output extraction. AI tools are used to accelerate specific tasks — not to generate content I haven't reviewed, make decisions I haven't made, or produce expertise I don't have.

Pages written by me are written by me. When AI assists, it assists with formatting and structuring, not with the underlying knowledge or judgment.

Transparency about limitations

I am one person, self-funded, working independently. I make mistakes. The resource databases may have outdated listings. The sites may have bugs. The tools may fail. I fix things when I find them and when people tell me about them.

If you find an error — in a resource listing, a site, or anything I've built — tell me. The contact form exists for that reason.