Truligon as an AI and product engineering lab.

Problem

Modern product teams need builders who can move through ambiguity: architecture, UX, backend services, AI tooling, deployment, and iteration. Truligon is the working environment I use to practice that full loop.

Requirements

The platform needed to support multiple public surfaces without turning every experiment into a separate one-off deploy. It also needed auth/admin tooling, learning pages, contact and resume surfaces, and enough operational documentation to recover routes when a live subdomain stopped resolving correctly.

Architecture And Implementation

• Subdomain application architecture for multiple product surfaces.
• TypeScript and Node.js server systems with Express routing and API modules.
• Admin interfaces, authentication flows, organization management, and email tooling.
• Learning content surfaces, platform pages, dashboard concepts, and UX-focused frontends.
• Cloudflare-oriented deployment, DNS, tunnel, and domain setup documentation.

Tradeoffs And Maintenance

I kept the platform pragmatic: static Vite pages for portfolio and learning content, Express for routing and APIs, Prisma/SQLite where durable data was needed, and Cloudflare tunnel rules for public subdomains. That keeps iteration fast while still exposing real production concerns like host headers, asset paths, build outputs, and route-specific fallbacks.

Testing And Reliability

I verify changes with TypeScript checks, frontend builds, route scans, generated HTML parse fixes, accessibility checks, and live-domain troubleshooting. Recent reliability work included correcting subdomain ingress rules, build-output asset paths, and generated learn-page markup that could break Vite.

Outcome

This work shows end-to-end ownership: defining requirements, choosing a workable architecture, implementing frontend and backend behavior, debugging production routing, documenting setup, and keeping a live platform coherent as it grows.