━━ HISTORY / WORK & PERSONAL PROJECTS
Frontend work at Bally’s Interactive. Personal blogs rebuilt across Hugo, Next.js, and React. A record of how I build, publish, and refine for the web.
2022.09 — 2023.12 / UNITED KINGDOM
I worked as a frontend developer at Bally’s Interactive. Jackpotjoy is a representative example of the kind of website I worked on.
Developed React and TypeScript marketing landing pages for gaming platforms serving the UK, Europe, and North America.
Designed reusable component structures for recurring brand campaigns to minimize duplicated code. Worked with Styled-components and Storybook to maintain the UI system.
Integrated REST and GraphQL APIs, implemented responsive layouts, and addressed cross-browser compatibility.
Collaborated in English with designers, planners, and backend engineers in a multinational team, shipping to production on short marketing release cycles.
SOURCE: OWNER’S RÉSUMÉ · BALLY’S EXPERIENCE
The public pages below illustrate product context. They do not identify my individual feature ownership or reproduce the site as it was during my employment.
PUBLIC PRODUCT / JACKPOTJOY
Category navigation and grouped game listings connect a broad catalogue to individual pages.
VIEW SOURCE PAGE ↗Login and registration entry points sit alongside campaign modules and linked offer terms.
VIEW SOURCE PAGE ↗A dedicated blog extends the product with articles and navigation into editorial content.
VIEW SOURCE PAGE ↗01 / 2024.10 — 2024.12
Hugo / Markdown / Sass / GitHub Pages
A Hugo blog using the nostyleplease theme, with daily, weekly, monthly, review, travel, and English/Korean writing sections. Local templates and Sass customize the reading experience.
Configured a Hugo build and GitHub Pages deployment workflow; refined list limits and browser presentation. This version kept publishing close to Markdown files.
02 / 2025.01 — 2025.08
Next.js 15 / MongoDB → Supabase / Plate / Tailwind / Fly.io
Built a Next.js site with Fly.io deployment configuration, first with a MongoDB/NestJS search backend and later Supabase-backed blog and dev views, plus persistent light/dark themes. The composer evolved into Plate-based editing with separate create, edit, and read-only views.
Explored database-backed authoring, structured editor values alongside Markdown, and a TanStack Query experiment marked as work in progress in the commit history.
03 / 2026.03 — 2026.09
React 19 / Vite / TanStack Router / Supabase / TipTap
The current repository uses a Vite application, file-based routing, a unified content table, and TipTap publishing tools. Its visual identity brings the home, content, and gears pages together.
The detailed log below follows this repository. Together, the three repositories show successive approaches to personal publishing; they do not establish a direct code or data migration between every version.
Search was a separate full-stack feature: Next.js search UI → /api/search → NestJS /search → MongoDB Atlas Search.
The NestJS backend added MongoDB-backed content CRUD through Mongoose.
minju25kim-nest / deb0efa ↗The backend added an Atlas Search aggregation over title, content, and keywords with fuzzy matching (maxEdits: 2). The Next.js /api/search endpoint forwarded queries to the backend’s /search endpoint on the same day.
minju25kim-nest / 9fc21e7 ↗minju25kim-next / f22bf1e ↗Refined the frontend search with debounced requests, a slash-key shortcut, arrow-key result selection, and outside-click dismissal.
minju25kim-next / 168c774 ↗A later remodel added a shared MongoClient connection and queried the dev collection for content counts, categories, views, and recent posts.
minju25kim-next / 09114aa ↗The Next.js repository removed its MongoDB helper and introduced Supabase-backed development content. This dates the code change; it does not establish when the separate backend stopped running.
minju25kim-next / 3f24931 ↗Set up React 19, TypeScript and Vite with React Compiler, file-based TanStack routing, and a Supabase client for blog data.
Established the application shell, typed routes, cached blog reads, and a Vitest test environment for subsequent features.
Added a daily GitHub Actions workflow to refresh a Strava OAuth token and sync yearly totals and recent activities into Supabase.
Connected a scheduled data pipeline to an activity page. The integration and its workflow were later removed on July 1.
Built a Supabase-backed development notes route, then added GitHub sign-in, protected editor routes, and a TipTap rich-text editor.
Moved authoring into the blog itself, with formatting controls, edit flows, and published-only post listings.
Built the gears page around Supabase records and Storage image URLs, with category chips, specifications, and scroll-driven entrances.
Extended the site beyond posts with a dedicated presentation for equipment and its details.
Merged blog and development posts into one content table, migrated existing data, and moved the reading and editing routes to /content.
Created a shared publishing flow with content types, UUIDs, published-read policies, and pagination in batches of 20.
Replaced the earlier green-accented interface with a newspaper-inspired brutalist design: monospace type, oversized headlines, thick rules, and orange highlights.
Brought the home, content, and gears pages into a shared visual language using reusable palette tokens and ledger-style layouts.
Tightened editor route checks and added an admin-only database write policy. Added slug generation, sanitization, uniqueness checks, and a draft/live management page.
Added confirmed deletion and admin draft previews while hiding unpublished posts from public readers.
Improved mobile navigation and responsive layouts, unified not-found screens, raised the desktop breakpoint to 1280px, and increased gear text sizes.
Gave narrow windows a single-column layout and replaced placeholder gear statistics with values derived from data.
Removed the fake project showcase from the site.
Kept the portfolio focused on work evidenced by the repository, rather than sample projects.
Removed the Strava route, sync script, scheduled workflow, and associated homepage activity statistics. The commit message cites API pricing as the reason.
Reduced the site’s integration footprint while preserving the implementation in Git history.
Deleted unused assets from the repository after earlier design and feature changes.
Continued the maintenance work behind the current blog.
Reviewed through . Blog chapters are curated from the three repositories’ commits, configuration, and source files. The dates describe their visible commit history, not guaranteed deployment periods.
The Bally’s dates and responsibilities are drawn from my supplied résumé; product observations link to current public Jackpotjoy pages. Several current-blog commits credit Claude as a co-author. Private ChatGPT conversations were not used.