Personal Portfolio
Production-grade bilingual portfolio shipped in 12 hours via AI-accelerated development and latest technologies, architecture optimized for performance, SEO, and scalability.
Next.js 15 architecture with SSR, full i18n, and MDX; p99 FCP < 200 ms, full dark mode support, and automatic scaling on Vercel.
Duration
Present●
Introduction
Launched a bilingual AI-accelerated portfolio in 12 hours with Next.js 15 + SSR, achieving Lighthouse >95 and FCP p99 <200 ms. Built for speed, scalability, and clarity, it's responsive, SEO-optimized, accessible, and production-grade—with zero post-launch incidents.
Shipped in 12 hours: clean architecture, i18n, MDX, and dark mode. AI accelerated repeatable tasks; I made the design/system calls to maximize Lighthouse and load times.
The Challenge
The main challenge was time: deliver a complete product from scratch, including internationalization, responsive UI, MDX content, and dark mode in under 12 hours. Success required ruthless prioritization and tight scope control. Some models delivered excellent architectural scaffolding; others had to be pushed hard to avoid bloated code or artificial-sounding text. The key was using AI not just for speed, but with precision.
Solution & Approach
The architecture was built on Next.js 15, with strict TypeScript, Tailwind for rapid UI development, and Vercel for CI/CD. I delegated repetitive tasks (boilerplate, markdown generation, copywriting) to AI, while keeping full control over structure, UX, and product logic. Speed was the differentiator: every section, every block of content, every component was generated, reviewed, and deployed in minutes.
Technical Highlights
- Next.js 15: Server-side rendering and SEO optimizations out-of-the-box.
- TypeScript (strict mode): Ensured reliability and safe refactors across the codebase.
- Tailwind CSS: Rapid responsive UI development with clean design system.
- Vercel: Zero-config deployments and automated CI/CD.
- next-intl: Fully managed i18n with locale routing and translation primitives.
- MDX: Dynamic content embedding with markdown simplicity.
Models Used
- Claude 4 Opus: Used for bootstrapping core architecture and generating complex components.
- Claude 4 Sonnet: Assisted with fine-tuning logic and making incremental improvements.
- ChatGPT 5: Helped define section structure and refine narrative flow.
- Gemini 2.5 Flash: Used to rapidly draft and iterate on content variants.
Results
Achieved Lighthouse scores >95, 100% uptime since launch, and zero post-launch bugs—showcasing that AI-boosted delivery can produce polished, resilient results in hours, not weeks.
Key Metrics
- Lighthouse scores >95 across performance, accessibility, and SEO
- FCP <200ms at 99th percentile
- 5x faster delivery vs traditional methods
- Zero post-launch issues, 100% uptime
This project proves how thoughtful use of AI, combined with clear architectural decisions, can dramatically accelerate product delivery without compromising quality. It sets a repeatable baseline for future prototypes and open source contributions, while doubling as a live, evolving portfolio of execution.