|
What it does: LusoPronounce teaches Brazilian Portuguese pronunciation through phoneme-level analysis using Azure Speech AI. It provides phoneme tips and recommendations via OpenAI / Google AI Studio APIs. Links: • 📂 GitHub Repo • 📄 Presentation Slides Tech Stack:  Azure Speech AI Next.js FastAPI MongoDB |
|
What it does: SynapsePRD turns product ideas into build-ready specs with AI. Users can refine requirements, explore alternate directions, generate UI mockups, and produce implementation artifacts from a single source of truth. Key Features: • AI-generated PRDs with branching for iteration and alternatives • UI mockup generation from wireframe to high-fidelity • Auto-generated artifacts like flows, screen inventories, data models, and implementation plans • Prompt generation for coding agents and developer workflows Links: • 📂 GitHub Repo • 🚀 Live Demo Tech Stack:  React 19 TypeScript Vite Tailwind CSS🤖 Google Gemini 2.5 (Flash/Pro) 🐻 Zustand (State Management) ▲ Vercel (Deployment + Serverless)  Playwright (E2E Testing) |
|
What it does: A platform for exploring advanced Monte Carlo Tree Search (MCTS) techniques in complex, multi-player environments. The project focuses on both building a strong agent and visualizing how MCTS behaves, including tree growth, exploration patterns, rollout distributions, and move selection probabilities. Links: • 📂 GitHub Repo • 📄 Presentation Slides Tech Stack:  Python (MCTS, Bitboards) FastAPI React / TypeScript Tailwind CSS |
|
What it does: A multi-mode activity execution system with checklist, image-guided, and text-guided flows, plus automated habit logging and MongoDB-backed tracking. Links: • 📂 GitHub Repo • 📄 Presentation Slides Tech Stack: MongoDB React TypeScript Node.js |
|
What it does: SpatialTasks is a local-first visual task system built on an interactive node canvas. Users lay out work as action and container nodes, connect them with dependency edges, and drill into containers for nested subgraphs with breadcrumb navigation—keeping projects spatial and relational instead of trapped in a flat list. Key Features: • Execution mode that highlights the next actionable work, dims finished or blocked items, and pans between steps • Step-detail panel inside containers with substeps, verification notes from imports, and complete-and-advance flow • List view for scanning and sorting tasks alongside the canvas • Resizable nodes, per-node notes, undo/redo, and browser persistence with save indicator • Markdown import from file drop or paste, with editable draft review before canvas creation • Optional Gemini (BYOK) for Magic Expand and AI-generated starter flows • Multiple projects, optional Supabase sync/auth, and touch-friendly UI patterns Links: • 📂 GitHub Repo • 🚀 Live Demo Tech Stack: React 18 TypeScript Vite🕸️ React Flow 🐻 Zustand + zundo Tailwind CSS🟢 Supabase (optional sync) 🤖 Gemini API (optional AI) Interactive Canvas UI: Pan/zoom graph, drag nodes, connect dependencies, nest workflows in containers, resize nodes, and switch to list view without leaving the project. |
|
What it does: A procedural harmonic generator and immersive music sketchpad. It constructs scale-aware chord progressions and melodies using algorithmic voice-leading heuristics, paired with a real-time Tone.js synthesis engine and interactive piano roll. Links: • 📂 GitHub Repo • 🚀 Live Demo • 📄 Presentation Slides Tech Stack: TypeScript Next.js Tailwind CSS🎹 Tone.js |
|
What it does: Generates ergonomic finger-drumming layouts with movement-cost optimization, RL exploration, and onion-skin gesture overlays. Links: • 📂 GitHub Repo • 📄 Presentation Slides Tech Stack: Optimization Engine React + SVG Node.js |
|
TypeScript |
Python |
PyTorch |
Scikit-learn |
|
Azure |
Git |
Docker |
CursorAI |
Google Antigravity |
React |
Node.js |
Pandas |
OpenCV |
MongoDB |
PostgreSQL |
Ableton Live |
🔭 Working on Generative AI for music, art, and learning
🌱 Studying for the Azure AI-102 Exam
✏️ Credential: Azure AI Natural Language Processing