Overview
An Ace Attorney inspired interactive portfolio chatbot where Miles Edgeworth presents evidence of professional qualifications through a courtroom-themed experience.
Scope
Full Stack Architecture, AI Integration, Interactive UI Design
Tech Stack
Languages: TypeScript, HTML/CSS
Frontend: React 19, Next.js 16 (App Router), Tailwind CSS, Framer Motion, Radix UI, Lucide React
Backend: Next.js API Routes, NVIDIA Inference API (MiniMax M2)
Data Layer: Local JSON (RAG-style evidence retrieval)
File Processing: Tesseract.js (OCR), unpdf, mammoth (PDF/DOCX parsing)
Developer Tools: Bun, ESLint, Git/GitHub, Turbopack
Challenge
The primary challenge was designing a persona-driven chatbot that maintains character consistency while delivering accurate, context-aware responses. This required implementing RAG-style evidence retrieval from local JSON files, integrating sentiment-aware prompt strategies with the NVIDIA Inference API, and handling diverse file uploads (OCR for images, PDF/DOCX parsing) without compromising response latency.
Objective
The objective was to create an unconventional, memorable portfolio experience that showcases technical skills through an engaging Ace Attorney–themed interface. The goal was to combine AI-powered conversation with RAG evidence retrieval, enabling users to explore professional qualifications through natural dialogue with the Miles Edgeworth persona.
Results
Miles Edgeworth — The Lawyer delivers a polished, immersive portfolio experience featuring animated sprite states, objection overlays, and sentiment-adaptive responses. The application supports file uploads with OCR, URL content ingestion, and quick-question chips—all wrapped in a 100% TypeScript codebase with secure server-side API handling.
