
An academic hub built for the ACECR AI Department. It streamlines research publication catalogs, showcases applied enterprise AI models, and provides a multi-step data submission pipeline with strict schema validation for academic research files.
Handling asynchronous, multi-format research dataset uploads while maintaining a fast initial load and clean UI animations on an academic audience's diverse hardware.
Built on Vite, Bun, and React 19 for minimal bundle overhead. Combined React Hook Form with Zod for zero-latency client validation and TanStack Query for asynchronous server synchronization.
Micro-bundled production output minimizing client parsing overhead.
Provides non-blocking, isolated field re-rendering during complex inputs.
Validates MIME signatures and size constraints directly on client threads.
Leveraged Bun and Vite SWC compiler to cut CI/CD compilation times.
Strict Zod schemas prevented malformed academic submissions before network transit.
Full keyboard navigation and ARIA landmarks verified via automated audits.
Academic institutional infrastructure had restricted server execution budgets; a static, CDN-cached SPA eliminated ongoing server hosting overhead while maximizing speed.
Prevented layout re-renders on every keystroke in multi-page document upload forms.
Interactive inspection of high-fidelity views
