These are student/volunteer case studies. They run locally or on small hosts, they lean on AI scaffolding, and each card explains what’s real versus still on the roadmap.
Want a deeper teardown? Reach out on LinkedIn or GitHub and I’ll walk through repos, TODO lists, and prompt logs so you can see the current state before assuming production polish.
Reality: Frontend runs on GitHub Pages and the backend spins up on Render & Atlas, but cold starts take ~5 minutes before login/chat/forums respond. Events, profiles, and forums do persist once the free instance wakes up, so I call it a demo rather than a production app.
Practice build that let me explore React forms, profile cards, and an Express API (MongoDB Atlas + Render) with help from tutorials + AI scaffolding.
Reality: Implements keyboard navigation and basic skeleton states. There’s no offline mode or advanced caching yet—this is purely a learning sandbox.
Simple client that consumes the public Pokémon API so I can keep practicing search, filtering, and responsive layout basics.
Reality: A few minor PRs merged; most work happens in forks while I learn. I’m transparent about scope so maintainers know where I’m still getting up to speed.
Student contributor making docs updates, small fixes, and experimental branches for the CIRIS Ethics Engine repos.
Reality: Focus management and reset hooks work in modern browsers; it’s a CodePen-style experiment, not a shipped classroom tool.
Front-end warm-up (built with AI prompt scaffolding) that let me practice keyboard handling, audio resets, and small delightful details.
Reality: Hosted at bradleymatera.github.io/Professional-Portfolio-Website/ as a demo that links into this Gatsby/Netlify portfolio (https://bradleysgatsbyblog.netlify.app). It’s a reference template, not a separate product.
Semantic template I built (with heavy AI help) to practice landmark structure, skip links, and responsive typography.
Reality: Primarily for my own study sessions. I link to specific pens from blog posts so readers know which snippets were AI-assisted.
UI experiment log where I stash pens generated with AI prompts + manual tweaks so I can reference patterns later.
Reality: Runs locally only; still uses AI scaffolding for most of the backend code. Great for learning infrastructure, not ready for users.
FastAPI + Ollama playground for testing local chat flows, env variables, and prompt templates on my laptop.
Reality: Currently supports a single endpoint and a handful of scripts. Analytics/guardrails/billing are ideas in the backlog, not shipped features, and I call out which parts were generated vs. hand-written in the README.
Work-in-progress embed that lets me experiment with prompting, simple Express proxies, and UI hand-offs for chat assistants (most scaffolding from Copilot/ChatGPT).
Reality: Powered entirely by mocked data and AI-generated copy. Useful for storytelling, not hooked to real metrics.
Static dashboard mock that let me test layout ideas for an internal analytics concept.
Reality: Built for my own onboarding, not adopted by teams. The README calls out what still needs polishing before anyone else could rely on it.
Personal exercise to learn how Docker Compose can spin up Node + Python tooling with similar scripts.
Reality: Lives at https://bradleymatera.github.io/CheeseMath-Jest-Tests/ as a front-end demo. The Jest repo is still a practice reference (not production-ready), but the live site lets folks try the calculator, regex tester, and card obscurer without cloning the code.
Dual demo: A GitHub Pages UI (Next.js 16 + NextUI) showcasing the calculator features described in the README, plus the Jest practice repo that backs the basic math utilities. Built with heavy AI help for the copy and components.
Reality: Handles small meshes today; larger files still surface TODOs in the README.
Zig learning repo where I parse OBJ/MTL files to understand memory management and file I/O outside of JavaScript.
Reality: Personal experiment, not a framework. It keeps the sample code, adds plain copy about adapters/buffers, and reminds visitors that this is a self-taught WebGPU study—not production graphics work.
Small Next.js + TypeScript rebuild of the Hello Triangle and Textured Cube demos so I could map device setup, pipelines, and WGSL shaders to a UI. I leaned on AI prompts and the official samples to keep the explanations accurate.
I’m happy to schedule a walkthrough, record a demo, or share repo access so you can see the code quality, test coverage (or lack thereof), and open TODOs for any project listed here.