Projects

Each project is framed as an AI-augmented engineering case study: architecture, delivery workflow, execution layers handled by agents, and the control points where engineering judgment still matters.

Featured Projects

  • PixelCollab: Real-Time Collaborative Whiteboard
    Featured build / live

    PixelCollab: Real-Time Collaborative Whiteboard

    PixelCollab is a multiplayer whiteboard built for fast visual collaboration. Users can create a room, share a link instantly, and draw together on a live synced canvas with presence indicators, chat, color tools, and a clean focus-first workspace.

    Frontend product design, realtime collaboration UX, and interaction polish

    • React
    • TypeScript
    • Canvas UI
    • Realtime sync
    • Presence indicators
    • Share-a-link room creation keeps collaboration friction low from the first action
    • The canvas, chat, and presence states are designed to stay readable in one workspace
    View details Live demo
  • AI Task Orchestrator: Autonomous Delivery Planning System
    Featured build / case study

    AI Task Orchestrator: Autonomous Delivery Planning System

    A project concept for turning feature requests into structured execution plans: task decomposition, AI-agent routing, implementation prompts, test suggestions, and delivery checkpoints.

    AI Engineer, system designer, orchestration workflow architect

    • Next.js
    • Node.js
    • OpenAI API
    • BullMQ
    • PostgreSQL
    • GitHub Actions
    • Designed to turn a feature brief into an executable delivery plan with architecture, tasks, tests, and PR structure
    • Separates orchestration, execution, and validation so AI agents can move faster without losing control
    View details

Open-source and Infrastructure

  • thuongdadev-v2: Portfolio Platform and Editorial Workflow
    Open source / live

    thuongdadev-v2: Portfolio Platform and Editorial Workflow

    A production-minded personal portfolio built to present work with stronger technical context: Astro SSR, structured content, metadata strategy, and a GitHub-backed CMS workflow.

    Product design, frontend architecture, content modeling, and deployment setup

    • Astro SSR
    • React
    • TypeScript
    • Tailwind CSS
    • Decap CMS
    • Docker
    • Structured content model for both articles and project case studies
    • SEO foundation includes canonical URLs, Open Graph data, Twitter cards, and JSON-LD
    View details Live site GitHub

Next AI-Native Builds

Instead of hiding unfinished work, I keep the next AI-native systems visible. These are the builds I want to turn into deeper case studies around autonomous delivery and engineering leverage.

Roadmap

AI Code Review Assistant

A PR analysis workflow that summarizes change risk, highlights refactor opportunities, and generates structured review output before human approval.

PR analysis, risk detection, refactor guidance, review automation

Roadmap

Autonomous Delivery Dashboard

A system for tracking agent-assisted feature delivery, code review status, testing progress, and deployment readiness across one delivery pipeline.

AI orchestration, delivery visibility, CI gating, execution control

Roadmap

Observability for AI Workflows

A case study around tracing prompt flows, task execution, and human-in-the-loop approval inside an AI-native engineering stack.

Prompt tracing, logs, QA checkpoints, workflow reliability