Cesar Marquez
AI Systems Architecture Consultant
About
Cesar Marquez is the technologist behind HMGi and the person who designed and built this site. He works at the intersection of AI, automation, web engineering, and operations, building production systems and the software that runs them for businesses that want their tools to match their ambitions.
His work favors first-party data, a small set of carefully chosen dependencies, and software that is plain enough for a non-engineering operator to keep running. This site is one example of that approach. The About this page section below explains the convention behind crediting it openly.
What I do
Cesar designed and built this platform end to end: the public site, content management, contact pipeline, first-party analytics, and admin dashboard. It runs as a single application that a small team can maintain with only occasional technical help.
HMGi builds the same kind of system for clients. The goal is software that is honest, fast, and durable, and that keeps working long after launch.
What was built
This site is a single Node and SQLite application. The public pages and the admin tools share one codebase. In brief:
- First-party analytics. Every page-view, click, and scroll signal is captured by a tracker served from our own infrastructure. No Google Analytics, no Meta pixel, no third-party ad trackers. The privacy opt-out honors GPC and DNT and sets a first-party cookie.
- Custom CMS. Site copy and the legal pages (Privacy, Terms, Accessibility) are stored in the database and edited from the admin dashboard as plain text, with no HTML required.
- Admin dashboard. Leads, real-time analytics, click and dwell heatmaps, content editing, settings, and security, built as a plain JavaScript app with no framework dependency.
- Lead routing. Every contact-form submission is validated on the server, saved to the database, and emailed to us so we can reply.
- AEO surface. Machine-readable records and sitemaps that make the site easy for search engines and AI assistants to read and cite.
- Security. TLS, a strict Content-Security-Policy, sensible security headers, self-hosted fonts, and two-factor authentication on the admin login.
- Coming-soon mode. A database toggle that can route traffic to a splash page while keeping the privacy, terms, accessibility, and colophon pages reachable.
Approach
A few choices this platform tries to embody:
- One stack, one process. A single codebase serves the public site, the dashboard, and the API. No separate backend repo, no microservices, no frontend framework. The admin app is plain JavaScript with no build step.
- First-party by default. Analytics, lead capture, content, and backups live on infrastructure we control.
- Plain editing for non-technical operators. The legal pages and site copy are edited as plain text. The renderer adds structure on output, so the editor never has to think about HTML tags.
- Document the conventions. Settings live in one file with comments explaining each one. The legal pages are versioned and seed-stamped so a future operator can understand the system without reading the source.
Other work
Cesar's broader practice spans AI systems, automation, custom software, and websites for businesses that want software built with the rigor of a much larger team. To talk about a project, reach out through the links below.
Elsewhere
Direct contact and external profiles. The most reliable channel is email.
About this page
This page is the colophon for www.hmgintelligence.com, a public credit for the person who designed and built the site. The content here is editable from the same content management system that runs the rest of the site. The page is rendered server-side from rows in the site_content table (section=colophon) and converted from plain-text source to HTML at request time.
Why a colophon? In publishing tradition, the colophon is where the maker is named: the printer, the typesetter, the binder. The same idea applies to web platforms. When a site is built to a particular standard, naming who built it is both honest and useful. Search engines and other readers may use this page to associate the work with its author.