You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
GSAP ScrollTrigger setup for fade-up entrance animations
useReducedMotion
Detect prefers-reduced-motion for accessibility
useMobileMenu
Mobile nav open/close state
Animation Implementation
Animation
Library
Approach
Complexity
Hero neural network canvas
Three.js
Raw Three.js: BufferGeometry points, custom shader material for glowing nodes, line segments for connections within threshold, mouse raycaster for attraction force. Drift via per-node velocity in render loop.
High
Hero text split reveal
GSAP
SplitType or manual span splitting, staggered gsap.from with y: 30, opacity: 0, duration: 0.8, stagger: 0.03
Medium
Page load sequence
GSAP
gsap.timeline() orchestrating nav → name → subtitle → stats → CTAs with calculated delays
No backend — form uses mailto: link or Formspree (client-side only, can use formspree.io endpoint if desired, otherwise just mailto fallback)
Scroll-Triggered Nav Background
useState<boolean> for scrolled
useEffect with scroll listener: window.scrollY > window.innerHeight * 0.8 → set scrolled
Apply bg-[#0A0E17]/90 backdrop-blur-xl class when scrolled
Other Decisions
Raw Three.js over React Three Fiber
The hero neural network is a single self-contained canvas with custom particle logic, mouse interaction, and connection rendering. Raw Three.js in a useEffect hook provides full control without R3F's declarative overhead. The canvas is isolated in NeuralCanvas.tsx.
GSAP over Framer Motion
GSAP is chosen for: (1) ScrollTrigger for scroll-driven animations, (2) timeline sequencing for the page load, (3) more control over stagger and easing for the split-text effect. Framer Motion's whileInView is simpler but less precise for the orchestrated load sequence.
No Backend
This is a static portfolio. Contact form uses a mailto: link or simple Formspree integration. No API routes needed.
Image Strategy
Project card images are generated AI visuals. Use object-fit: cover in the card image area. Implement lazy loading with loading="lazy" and a dark placeholder background.
shadcn/ui Usage
Minimal use of shadcn/ui components — this is a highly custom portfolio. Only Button and Badge variants may be used if they fit; otherwise custom components for full design control.