Building Scalable Products, Intelligent Systems.
I deliver full-stack and AI-focused projects using Nuxt 4, Nuxt UI, Tailwind, FastAPI, Docker, and PyTorch. Here are real examples of architecture, UX, and ML integration.
2025
AI Activity Planner — Grok-Powered Assistant
AI-powered activity planning platform using xAI's Grok API to recommend outfits and activities based on real-time weather and location data. Built with Python 3.12 (FastAPI) async backend and Nuxt 3 (Vue.js) with full TypeScript support for type-safe frontend development. Backend integrates OpenWeatherMap, Google Geocoding, and FourSquare/Google Places APIs with Tortoise ORM + MySQL, Redis 7 caching, and Celery task queues. Features include real-time weather dashboard, personalized outfit suggestions, smart activity curation, auto-generated packing checklists, multi-day itinerary planning, conversational interface with Grok, weather alerts, and social sharing. Frontend uses Pinia (TypeScript) for state management, Tailwind CSS with shadcn/vue, VueUse composables, and Framer Motion animations. Recommendation system includes user feedback loop for continuous prompt optimization. Containerized with Docker, deployed on AWS ECS with CI/CD via GitHub Actions, and monitored with Prometheus + Grafana + Sentry. Achieves 90%+ test coverage with Pytest and Vitest.
2025
Enterprise Digital Showcase — FastAPI + Nuxt 3
Enterprise-ready application built with Python 3.12 (FastAPI) and Nuxt 3 (Vue 3). Backend features asynchronous request handling via Uvicorn, Tortoise ORM for async MySQL operations, JWT authentication with refresh tokens, RBAC permissions, rate limiting via SlowAPI, and Celery + Redis for background tasks. Frontend leverages Nuxt 3 SSR/SSG, Pinia for state management, and Tailwind CSS for responsive UI. Core features include project portfolio CRUD, full-text search with faceted filtering, real-time dashboards with Chart.js, asset management with Cloudinary, client collaboration portal, WebSocket notifications via Socket.IO (Redis-backed), and PDF/Excel report generation. Redis 7 provides distributed caching and session storage. Containerized with Docker, deployed on AWS ECS with auto-scaling, CI/CD via GitHub Actions, and monitoring with Prometheus + Grafana. Achieves 90%+ test coverage with Pytest and implements Helmet.js for security hardening.
2025
Food Lovers — Recipe Platform with 1,000+ Recipes
Food Lovers is a full-stack culinary platform built with Python (FastAPI) and Vue 3. The backend uses FastAPI with asynchronous request handling, Tortoise ORM for async MySQL operations, and JWT authentication with refresh tokens. Redis caching and Celery task queues power background jobs (image processing, email newsletters). The frontend leverages Vue 3 Composition API, Pinia for state management, and Tailwind CSS for responsive design. Features include interactive step-by-step cooking guides with timers, full-text recipe search with Elasticsearch, user-generated content (recipes, ratings, comments), auto-generated shopping lists, nutritional insights via Open Food Facts API, WebSocket real-time sync, and weekly email recipe drops. Containerized with Docker and deployed on AWS EC2 with Nginx reverse proxy. Includes 1,000+ curated recipes with 250+ authentic East African dishes and a community of 500+ food enthusiasts.
2025
Injustify/EZED — Universal Media Downloader (V1 → V2)
Injustify/EZED is a cross-platform media downloader supporting YouTube, TikTok, Facebook, Instagram, and 1,800+ platforms via yt-dlp. The project evolved from V1 (Flask + SQLite + basic Vue) to V2 (FastAPI + MySQL + Vue 3 TypeScript + Nuxt UI + Celery/Redis), featuring AI-powered lyrics generation (Genius API + LRClib), audio separation (Spleeter), video summarization (OpenAI/Gemini), and real-time WebSocket progress tracking. V2 includes batch/playlist downloads, live stream capture, SponsorBlock integration, 6 customizable themes, browser extension bridge, and full Docker containerization. Backend uses Tortoise ORM + MySQL with Redis caching, Celery for async processing, and JWT authentication. Frontend uses Pinia state management, Tailwind CSS, and Nuxt UI components. Deployed on AWS ECS with CI/CD via GitHub Actions.
1970
Internal Developer Hub
2025
KenMall X — Full-Stack E-Commerce Platform
KenMall X is a full-stack e-commerce platform built with Next.js 14 (App Router) and React 18. It features SSR, SSG, and ISR for performance and SEO. State is managed with Zustand and React Query, while Tailwind CSS + shadcn/ui deliver a responsive UI. Backend uses Next.js API Routes with JWT authentication (HTTP-only cookies), MongoDB + Mongoose for data modeling, and Socket.io for real-time updates. Stripe Checkout handles payment simulation, MongoDB Atlas Search enables product discovery, and the admin dashboard uses Recharts. Dockerized and deployed on Vercel.
2025
Pothole Spotter — AI Road Defect Detection
AI-powered web application detecting, classifying, and visualizing potholes on Kenyan roads using Azure Custom Vision computer vision models (90%+ accuracy). Users upload road images with GPS coordinates, receiving real-time severity classification (minor/moderate/severe). Built with React 18 and Leaflet.js for interactive map visualization with color-coded markers and cluster grouping. Backend uses Node.js + Express.js with Multer for image uploads, JWT authentication, and BullMQ + Redis for async AI processing queues. MySQL database with geospatial extensions stores detection metadata and supports proximity searches. Features include road stretch analysis, admin dashboard with moderation tools, image compression via Sharp.js, CDN delivery, real-time WebSocket notifications, and CSV/PDF reporting. Redis 7 provides caching and session management. Containerized with Docker, deployed on Azure/AWS with CI/CD via GitHub Actions.
2025
SMS Sender — Bulk Messaging Tool
Bulk SMS messaging tool built with Python 3.12 (FastAPI) backend and Vue 3 (TypeScript) frontend with Nuxt UI. Enables users to send SMS messages to multiple contacts simultaneously with automatic SMS segment calculation (GSM-7/UCS-2 encoding). Features include contact management with country codes, message templates, bulk sending, scheduled delivery, real-time progress tracking, and delivery status webhooks. Integrates multiple SMS gateways (Twilio, Africa's Talking, Vonage) with automatic failover. Backend uses Tortoise ORM + MySQL, Redis 7 for caching and Celery for async processing. Frontend uses Pinia (TypeScript) for state management, Tailwind CSS, and VueUse composables. Includes SMS cost estimation, message history with analytics, CSV contact import/export, duplicate detection, and automatic retry logic. Containerized with Docker, deployed on AWS ECS with CI/CD via GitHub Actions, and monitored with Prometheus + Grafana.
