Justin K Lee
Full‑Stack Engineer with more than four years of experience building scalable internal business systems in agile, collaborative team environments. Eager about AI and emerging tech and always exploring what’s next. Nature hiker, museum wanderer, and lifelong learner at heart.
Who I am
Full Stack Developer I’m a Full‑Stack Developer with more than four years of professional experience contributing to scalable internal business systems as part of agile, cross-functional teams. I enjoy solving complex problems with clean, efficient code. My focus has evolved toward developing AI-powered applications, particularly those integrating Large Language Models (LLMs) into real-world systems. I design and implement LLM-driven features, automation workflows, and intelligent interfaces that transform traditional applications into adaptive, AI-enhanced platforms.
Outside of work, I recharge by spending time with family, hiking scenic trails, and visiting museums wherever I travel. These moments help me stay curious, grounded, and creatively inspired.
Main Skill Set
TOOLS
Works
Travelers to Canada AI
A fully interactive data visualization dashboard built with TypeScript, D3.js, Three.js, and browser running custom LLM model, designed to showcase 20 years of Canadian tourism statistics within a single, responsive web page. The app combines modern web standards—modular TypeScript architecture, CSS Grid and Flexbox, scalable SVG graphics—with GPU-accelerated 3D (Three.js: idle plane and province hover fly-in, non-blocking on the main thread) and AI-powered chat so users can ask questions about the data and get answers even when offline.
LLM and prompting are central to the experience: when online, the Groq API answers with rich, contextual responses; when offline or as a fallback, a mini LLM runs entirely in the browser using transformers.js (no server required, model cached after first load). Careful prompting—system instructions, dataset context injected by year/month, and routing of casual vs. data questions—ensures the assistant uses only the app’s filtered dataset for statistics and stays friendly for greetings and follow-ups. Technical highlights include advanced D3.js charting (threshold, linear, and square-root scales), real-time data filtering, a Three.js 3D plane layer (GLB model, instanced fly-in to provinces), and a maintainable codebase with strict linting and strong type safety.
Share Desk
A full‑stack workspace rental platform that allows users to list and book private offices and desk spaces around the world. Key features include geolocation-based search with Google Maps API, real-time availability, secure payment flows, and user authentication via Google and Facebook OAuth. The platform also supports listing management, messaging, and a user rating system.
Built with Ruby on Rails 7.1, Stimulus.js, PostgreSQL, Bootstrap 5, and deployed on Render.com. Google Maps and Geocoding API for interactive map search. Media uploads are handled via AWS S3, and the interface includes infinite scrolling, advanced search filters, and responsive UI components. JQuery for dynamic front-end behavior. Authentication is powered by Devise and OmniAuth, with API fallback logic and robust client/server-side error handling ensuring stability.
Collectors
A fully responsive, single-page e‑commerce application designed for premium movie and media collectors. Users can browse categorized products, manage shopping carts, and complete secure checkouts. The app integrates with Stripe for payment processing and Firebase for user authentication and backend data management, including support for Google OAuth login.
Built with React 19, Redux Toolkit, and TypeScript, the app uses Thunk middleware for managing async data flows and React Router for client-side routing. Global state is optimized through selector memoization, state normalization, and modular slice-based architecture. The UI is enhanced with code-splitting using React Lazy and Suspense, and performance is monitored via React DevTools profiling.
Key features include: dynamic product catalog with filtering and sorting, persistent shopping cart with local/session storage fallback, Stripe-integrated checkout flow with real-time validation, and deployment-ready setup with Vite
Get in Touch
Leave me a message