Arvince Ramos

Building a fitness web application with Next.js, TypeScript, Tailwind CSS, and Supabase. Implemented features such as a calorie and macro calculator, meal planner with a large curated food list, workout plans, and progress tracking. Wrote comprehensive Vitest test coverage and set up a GitHub Actions CI pipeline (typecheck, lint, test, build) with staging deployments on Vercel.

Developed a call-assistant web application using React and Node.js to support real-time notes and quick lookups during customer phone calls. Added autocomplete search to speed up and reduce spelling errors when entering dog breed names. Implemented automatic mapping of each breed to coat type and company price, reducing manual lookup work. Stored client profiles, call history, and pet notes in a database and supported ongoing updates to pricing and data.

Led a capstone project and built an AI voice travel assistant for both web and mobile. Designed and implemented a REST API with Node.js/Express running on Firebase Cloud Functions, using Firestore for data and Firebase Authentication and Cloud Messaging for user connectivity. Integrated OpenAI for chatbot capabilities, and used Serper and Geoapify for place data, with Google Maps Platform as a fallback for places, geocoding, and routing. Delivered voice mode with speech-to-text and text-to-speech, including itineraries, routing, weather, and reminders. Reduced API costs by switching to Gemini, local Ollama, and free OpenStreetMap services so the app could run on free tiers, while also securing user data with per-user Firestore rules, server-side token checks, rate limiting, and sanitized AI output.