Hi, I'm Michael Perez
A passionate web developer creating modern, responsive, and user-friendly applications. I specialize in React, Next.js, and TypeScript.
What I Do
I create digital experiences that are beautiful, functional, and user-centered.
Frontend Development
Building responsive, interactive user interfaces with React, Next.js, and modern CSS frameworks.
Backend & API Integration
Developing robust backend APIs, integrating databases, and ensuring seamless data flow for full stack solutions.
Mobile-First & Responsive Design
Designing applications that look and perform great on all devices, prioritizing mobile usability and accessibility.
Recent Projects
Here are some of my latest projects that showcase my skills and passion for development.
Henryetta Eye Center
A modern, responsive website for Henryetta Eye Center, a local optometry practice. Built with RNext.js, React.js and Tailwind CSS, featuring an appointment booking system and patient resources.
React.jsNext.jsLucideTailwind CSSVercelGoogle Maps API
Clarity-Ecommerce Framework / Clarity Payhub
A full-stack e-commerce application originally built with React.js and Bootstrap, featuring user authentication, payment processing, and admin dashboard. The project was then converted to Remix.js with Tailwind CSS.
Remix.jsTypeScriptC#MSSQLradix-uiTailwind CSS
Destination Athlete
Destination Athlete offers custom team gear, fundraising tools, and performance services for athletes and organizations. Built with ASP.NET, it uses JavaScript, jQuery, and Microsoft IIS for backend and hosting.
DNNAngularJSTypeScriptMSSQLBootstrap
Cool Projects
A selection of fun and interesting projects I've worked on in my spare time.
Month & Year Progress Bar Script
A small Bash script that prints 50-character progress bars and numeric stats for the current month, business days in the month, and the year (minutes/hours/days/weeks elapsed). Includes a watch mode to auto-refresh and requires GNU date / awk for accurate date arithmetic.
shell scriptingbash
Bit Flipper
A Python web app that visualizes numbers as binary bits grouped into 8-bit bytes with neon styling and animations. Features two modes: bit count to bytes conversion and binary bytes representation. Includes keyboard shortcuts for quick navigation and E2E tests with Playwright.
PythonFlaskJavaScriptCSSHTMLPlaywright
Pumpkin — Raspberry Pi Jack-o'-Lantern
A motion-sensing Halloween prop that rotates to face people who walk by. Built with a Raspberry Pi, PIR motion sensors, an MG995 servo and LEDs on a 3D-printed base.
See coverage: PiShop and Tom's Hardware.
See coverage: PiShop and Tom's Hardware.
Raspberry PiPIR motion sensorsServo (MG995)LEDs3D printing
AI-assisted Projects
Projects built with AI-assisted workflows, including autonomous local model orchestration and agent-driven development.
World News Globe
An interactive globe visualization that displays news articles from around the world. Users can rotate the globe to explore news by country, with articles categorized by topic and source. Built with React, Cesium, and various news API's and RSS feeds from around the world for real-time data. 3D globe visualization with clustered event pins. Multiple data layers: news, earthquakes, fires, weather alerts, conflicts, health, humanitarian. Live feeds from GDELT, NOAA, NASA, ReliefWeb, GDACS, WHO, and RSS sources. Search + geofencing: type a place & the camera flies there. Filter by severity / time window / keyword. Interactive overlay cards showing event details. Trends dashboard: keyword trends, sentiment, and geo heatmaps. Server-side caching via Redis (Upstash)
Next.jsReact.jsCesiumNews APIRSS
Todo App built with OpenClaw AI Agent and Local LM Studio
Built with React and TypeScript on a lightning-fast Vite stack, this todo app leverages Zustand for state persistence, Tailwind CSS for pixel-perfect styling, Lucide Icons for crisp UI, and Vitest for robust test coverage. The full experience is orchestrated by an autonomous OpenClaw AI agent using Local LM Studio, showcasing agent-driven development, AI-assisted workflows, and secure networking via Tailscale. The local model runs on a Tailnet-connected device, with medina-qwen3.5-27b-openclaw as the primary local LLM, unsloth/gpt-oss-20b as a backup sub-agent for cases Medina can’t resolve, and gpt-5-mini as a fallback before switching back to local inference to demonstrate the real capabilities of an offline-first local model.
ReactTypeScriptViteZustandTailwind CSSLucide IconsLocal LM StudioOpenClaw AI agentAgent-driven developmentTailscale