Lovefy
– AI Dating Platform
Jan 2026 — Mar 2026
Scroll to explore
Lovefy needed a high-performance web platform for AI-assisted, long-term dating — real-time messaging, photo sharing, fast page transitions, and AI-powered match assistance, all under real production traffic.
I joined as a Front-End Software Engineer acting full-stack, owning the real-time chat architecture end to end: building a Node.js + NestJS BFF, resolving backend API bottlenecks, and designing a hybrid rendering strategy (ISG + SSR + virtualization) to keep the app fast under load. I built the Next.js application on Vercel with React, TypeScript, Tailwind CSS, Shadcn UI, and React Query, and set up a CI/CD pipeline alongside AI agent workflows (Cursor and Claude) to speed up day-to-day development.
Highlights
- Built a Node.js + NestJS BFF layer that streamlined API contracts and resolved backend bottlenecks
- Implemented list virtualization and caching for instant, low-latency chat rendering
- Combined ISG for dashboards with SSR for fast, seamless page transitions
- Set up a full CI/CD pipeline (Vitest, RTL, Playwright E2E, Storybook) reaching 100% test and documentation coverage
- Introduced AI agent workflows (Cursor & Claude) that cut feature delivery time from 2 weeks to 4 days
Challenges
- Eliminating real-time chat latency without sacrificing UI smoothness, solved via SSR-aware list virtualization
- Owning full-stack scope as a front-end engineer, including designing and building the BFF layer from scratch
- Keeping bundle size in check while continuously shipping new AI-assisted features
- Speeding up delivery without cutting corners on test coverage or documentation
CompletedJan 2026 — Mar 2026
Next.jsNode.jsNestJSBFFVercelReactTypeScriptCursorClaudeTailwind CSSShadcn UIFigmaReact QueryReact VirtualizationVitestRTLPlaywrightStorybookCI/CDGitHub Actions