Executive Summary
A Filipino conversational voice and video AI study companion engineered for board and licensure exam examinees, delivering supportive active-recall tutoring in Taglish via Agora RTC.
"Board exam reviewees in the Philippines face high review center tuition costs, study burnout, and lack of interactive active-recall practice partners."
Cloud Architecture & System Design
Decoupled architecture combining a React (Vite) and Tailwind CSS frontend with a Node.js/Express backend, orchestrating Agora RTC streaming, Gemini 2.0 Flash reasoning, MiniMax Filipino TTS voice synthesis, and video avatar streaming.
Engineering Highlights
Spearheaded team engineering and prompt design for a conversational voice and video tutor submitted to the Agora Voice AI Hackathon Manila 2026.
Architected a multi-vendor AI pipeline integrating Gemini 2.0 Flash for domain reasoning, MiniMax for specialized Filipino TTS voice synthesis, and Agora RTC for low-latency bidirectional audio streaming.
Engineered a high-fidelity React and Tailwind CSS web application featuring conversational active-recall sessions, glassmorphism dashboard interfaces, and real-time study streak tracking.
Key Features
- Real-Time Conversational Voice & Video Tutoring via Agora RTC
- Culturally Tuned Taglish 'Ate/Kuya' Supportive Tutor Persona
- Multi-Model Pipeline (Gemini 2.0 Flash + MiniMax TTS + HeyGen/Anam Avatar)
- Active-Recall Review Session Dashboard & Topic Proficiency Metrics