
FULLSTACK AI LAB · SPEC-DRIVEN EDTECH · WASM & AI INTERVIEW ENGINE
Fullstack AI Lab
Ecosistema educativo interactivo de ingeniería de nivel empresarial basado en la Teoría de Carga Cognitiva de Sweller y el currículo en espiral de Bruner. Diseñado con Next.js 16 (React 19 Server Components, Actions y View Transitions), Tailwind CSS v4 y una canalización de caché en memoria que compila más de 1.700 páginas estáticas en menos de 90 segundos. Cuenta con un simulador de entrevistas técnicas con IA evaluadas con rúbricas Senior, sandbox WebAssembly de WordPress 7.0 / PHP 8.4 en el navegador sobre SQLite VFS, visualizadores <ArchitectureFlow>, simuladores de hooks y fallos en producción, evaluaciones formativas (+15 XP), tarjetas de repetición espaciada de Leitner, sincronización en la nube con Supabase Auth y plan PRO, indexación para agentes /llms.txt y prompts de audio para Google NotebookLM en 4 idiomas.
98+
PageSpeed
<1s
Page load
1,135+
MDX Lessons
225/225
Tests Passing
Frontend:
- Next.js 16.3 (App Router, Turbopack, and @next/mdx with native Rust mdxRs enabled)
- React 19.2 - Server Components (RSC), Server Actions, and <ViewTransition>
- TypeScript 5.8 - strict type safety across all components and API handlers with zero any tolerance
- Tailwind CSS v4 with @theme design tokens, CSS variables, dark-mode glassmorphism, and tw-animate-css
- Client-side WebAssembly environment running PHP 8.4 and SQLite VFS inside browser memory
- Zustand v5 state management with localStorage persistence and custom useIsMounted hydration guard
- Lucide React iconography and accessible UI primitives with responsive bottom navigation
- Node.js 24 LTS runtime environment for high-performance builds and pre-render pipelines
Backend:
- Edge and Node.js Serverless Route Handlers (/api/interview, /api/progress, /api/notebooklm, /api/revalidate)
- Supabase managed PostgreSQL database with Row Level Security (RLS) for cloud progress synchronization and PRO tier verification
- Automated technical interview scoring engine evaluating candidate code against Senior architectural rubrics
- Custom high-speed in-memory compilation and caching layer reducing SSG build time from 21.6m to 1.5m
API:
- Serverless route evaluating candidate responses, generating real-time feedback, and calculating readiness scores
- Declarative JSON Blueprint protocol mounting must-use plugins and configuring headless WordPress instances in WASM
- NotebookLM customized audio prompt pipeline generating first-principles podcasts in 4 languages
- Google OAuth and session management with Row Level Security for cross-device cloud progress sync
- Dynamic lesson metadata and localized search index generator across 1,135+ MDX documents
Gestión de contenidos (CMS):
- High-density file-based MDX architecture with 1,135+ lessons compiled via native Rust MDX-RS parser
- Pre-build AST scripts generating module trees, lesson URL maps, and multilingual navigation graphs
DevOps y seguridad:
- Automated build pipeline running TypeScript typecheck, Vitest test suites, and Rust content validation
- Automated production deployment with Global Edge Network, SSG pre-rendering, and instant rollbacks
- Automated pre-build content validator (scripts/validate-content.mjs) ensuring 100% Rust compiler safety across 1,135+ files
- Cloudflare edge proxy with SSL/TLS encryption, DDoS protection, and hardened HTTP response headers
Analítica:
- Local and cloud-synced tracking of lesson completions, quiz scores, daily streaks, and PRO tier progression
- GA4 integration tracking user journeys, module completion rates, and learning retention funnels
Herramientas de IA:
- Autonomous technical interviewer providing real-time architectural evaluations and Senior gap analyses
- Spec-Driven Development (SDD) methodology using Antigravity AI harness, modular rules, and quality hooks
- Zero-metaphor audio prompt synthesizer transforming MDX structures into deep-dive technical dialogues
- Used for rigorous architectural stress-testing, didactic spiral curriculum design, and TDD test authoring
SEO:
- Dedicated /llms.txt and /llms-full.txt endpoints delivering curated markdown context for autonomous AI search engines
- Next.js generateMetadata() generating OpenGraph, Twitter Cards, canonical URLs, and schema markup per lesson
- Complete 4-locale hreflang alternate headers (EN, RU, UK, ES) ensuring accurate multi-region Google indexing
- Automated XML sitemap with 1,700+ static URLs, canonical routing, and programmatic robots.ts
- Rich structured data (TechArticle, Course, LearningResource, BreadcrumbList) optimized for Google Rich Snippets and AI Overviews
Optimización técnica:
- Engineered for Core Web Vitals budgets: INP < 200ms, LCP < 1.2s, CLS < 0.05, sub-second TTFB
- 1,700+ static pages prerendered at build time with in-memory caching for near-instant navigation
- Custom useIsMounted hook eliminating React 19 hydration flickers for user-specific XP, auth, and streak state
- Automated <Term> interactive tooltips and <KeyPoint> callouts minimizing cognitive load during reading
Marketing digital:
- Didactics built on Sweller's Cognitive Load Theory: algorithm first (<StepFlow>), syntax second, formative testing third
- Retention mechanics: XP rewards, mastery level badges, daily streak counts, and PRO upgrade entitlements
- First-principles podcast audio overviews enabling hands-free commute and mobile learning without screen fatigue
Proyectos y casos de estudio relacionados
Explora otros proyectos con tecnología y objetivos similares

Unitrition
Plataforma de nutrición full-stack para encontrar alimentos adecuados según planes dietéticos específicos: elija un plan, busque alimentos y obtenga comparación instantánea por colores. Desarrollada con Headless WordPress, Next.js 16 (React 19) App Router, Supabase Auth y Calculadora de Unidades de Pan con API USDA FoodData Central. Incluye 10 dietas, soporte bilingüe EN/ES y blog educativo de nutrición con Schema.org.
Next.js 16 & Hybrid RAG Portfolio Platform
Plataforma fullstack de portfolio de nivel empresarial construida con Next.js 16 App Router, React 19 Server Components, TypeScript 6 y un asistente de IA Hybrid RAG propio que indexa 70 casos de estudio en 519 rutas estáticas localizadas (EN/UA/ES/DE). Incorpora telemetría GA4 con Consent Mode v2 Basic, integración con Google Search Console y distribución edge en Vercel y Cloudflare con accesibilidad WCAG 2.2 AA.
My React Portfolio Website
Aplicación SPA de portfolio de alto rendimiento construida con React 19, TypeScript 6, Tailwind CSS v4 y motor i18n trilingüe (EN/UA/ES), impulsada por la metodología AI-Harness 2026 (SDD + SDLC + TDD) con 29 specs de Vitest, 11 pruebas E2E en Playwright (WCAG 2.2 AA), Consent Mode v2 Basic y reducción del 50.3% del bundle inicial.