This is a submission for the New Year, New You Portfolio Challenge Presented by Google AI
About Me
I'm Hoang Manh Cam, a Senior Product Engineer who believes resumes should be an experience, not a document. In a world of infinite feeds and 3-second attention spans, why force recruiters to read a PDF?
My goal: Hijack the dopamine loop. I built a portfolio that feels less like a hiring portal and more like a high-end streaming service.
Portfolio
(Note: The projects featured in the video feed are **Technical Concepts* and MVP Showcases designed to demonstrate specific engineering challenges and solutions.)*
How I Built It
I engineered DevReels to mimic the "native app" fluidity of modern social platforms, but with the robustness of a production-grade web app.
- Core Engine: Next.js 14 (App Router) & TypeScript driving a highly optimized, client-side heavy interface.
- Visuals: TailwindCSS v4 for that sleek, "Void" aesthetic—using deep blacks (
#0A0A0A) and subtle gradients instead of flat greys to create depth. - Performance Physics: Powered by Framer Motion. I manually calculated snap points and scroll velocities to ensure that "satisfying" mechanical feel.
- The "Architect" (Google AI): Integrated Gemini 1.5 Flash to create a context-aware AI agents. It doesn't just chat; it sees what you're seeing. It knows the specific tech stack of the project on screen and answers questions like a Senior Engineer during a code review.
What I'm Most Proud Of
1. The "Cinematic" UX & Performance 🎞️
I ditched the standard "Infinite Scroll" for a more deliberate, Cinematic Feed.
- Smart Pre-loading: Videos rely on a custom
isVisibleheuristic. Only the active reel and its neighbors are rendered, preserving GPU resources while ensuring instant playback. - "System" Aesthetics: Instead of generic loading spinners, I built technical, data-driven placeholders that make even the waiting feel like part of the interface.
2. "Ask The Architect" - Contextual AI 🧠
Most portfolios have a generic chatbot. Mine has a Project-Aware Architect.
- Server-Side Security: All context data is validated server-side.
- Rate Limiting: Custom-built, in-memory rate limiting shields the Gemini API from abuse.
- It turns a static "About" section into a dynamic, 2-way technical interview. You can grill the AI about my code choices!
3. Obsessive Micro-Interactions ✨
The difference between "good" and "premium" is in the pixels.
- Responsive Drawers: On mobile, the UI morphs into touch-friendly Overlays. On desktop, it expands into a persistent Command Center.
- Glow Effects: Status indicators (like "Deployed • 2024") pulse with a radioactive green glow.
- Custom Cursors: The entire app uses custom cursor states to guide the user's interaction model subconsciously.

Top comments (0)