Skip to content
View PracticalSwan's full-sized avatar
3 followers · 10 following

Achievements

Achievement: Pull Sharkx2Achievement: YOLOAchievement: Pair Extraordinaire

Achievements

Achievement: Pull Sharkx2Achievement: YOLOAchievement: Pair Extraordinaire

Highlights

  • Pro

Block or report PracticalSwan

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
PracticalSwan/README.md

Sithu Win San

Computer Science student at Assumption University of Thailand
AI, machine learning, computer vision, recommender systems, and practical software engineering

Email GitHub LinkedIn

GitHub followers Profile views

About Me

I am a B.Sc. Computer Science student building applied AI and software projects with an emphasis on evidence, privacy, testing, and practical usability. My current work spans evidence-first AI research systems, recommender systems, local document intelligence, browser applications, and AI developer tooling.

I learn primarily by building complete systems: designing the architecture, implementing the product, testing failure cases, evaluating model or recommendation quality, and documenting the trade-offs.

Featured Projects

Project What it demonstrates
UniProof Evidence-first AI university research, comparison, and application guidance with Next.js 16, React 19, TypeScript, Supabase/PostgreSQL RLS, structured multi-provider AI, Vitest, and Playwright
Groovehaus Frontend + Backend Full-stack vinyl recommender system with React/Vite, Next.js, MongoDB/Mongoose, explainable recommendations, personalization controls, and offline recommender evaluation
OCR Model Local document intelligence using PaddleOCR, LayoutXLM, PyTorch, OpenCV, schema-validated extraction, Docker, and an optional privacy-gated Codex/MCP review workflow
Codex Router Local model-routing and Codex integration work around Responses API compatibility, provider adapters, credential isolation, Windows reliability, and model-catalog integration
Codex ChatGPT Control Plus TypeScript and Python SDK/tooling for visible, user-directed ChatGPT workflows through Codex, browser control, plugins, contracts, and release validation
Rubik's Cube Standalone Three.js/WebGL Rubik's Cube with manual interaction, solver-assisted recovery, Vite, Vitest, and workflow verification
Agent Skills Reusable AI-agent skills, validation workflows, and cross-client developer tooling

Current Focus

  • Evidence-grounded AI systems with deterministic validation and explicit uncertainty.
  • Machine learning, computer vision, OCR, and document information extraction.
  • Recommender systems, offline evaluation, ranking, personalization, and decision support.
  • Full-stack TypeScript/JavaScript applications with strong testing and security boundaries.
  • AI developer tooling using Codex, ChatGPT, MCP, model routing, and agent workflows.

Tech Stack

Languages

Python TypeScript JavaScript Java PHP SQL SML

Web and Application Development

React Next.js Vite Tailwind CSS shadcn/ui Radix UI Three.js Flask

AI, Machine Learning, and Data

PyTorch Transformers OpenCV PaddleOCR LayoutXLM scikit-learn Pandas NumPy Jupyter

Data, Backend, and Validation

Node.js MongoDB Mongoose Supabase PostgreSQL MySQL Zod

Testing, Deployment, and Developer Tools

Git GitHub Docker Playwright Vitest Netlify Vercel AWS Figma Blender

AI and agent tooling: OpenAI Codex, ChatGPT, Model Context Protocol (MCP), agent skills, model routing, browser automation, GitHub Actions

Current Semester Coursework

  • CSX3009 - Algorithm Design
  • CSX4207 - Decision Support and Recommendation System
  • CSX3004 - Programming Languages
  • CSX4213 - Computer Vision
  • CSX4201 - Artificial Intelligence Concepts
  • CSX4203 - Machine Learning
  • BBA1004 - Essential Marketing for Entrepreneurs
Selected completed coursework
  • CSX2009 - Cloud Computing
  • CSX3005 - Computer Networks
  • CSX3006 - Database Systems
  • CSX4107 - Web Application Development
  • ITX2007 - Data Science
  • CSX2008 - Mathematics Foundation for Computer Science
  • CSX3001 - Fundamentals of Computer Programming
  • CSX3002 - Object-Oriented Concepts and Programming
  • CSX3003 - Data Structures and Algorithms
  • ITX2004 - UI/UX Design and Prototyping
  • ITX2005 - Design Thinking
  • ITX3007 - Software Engineering
  • ITX3003 / CSX4401 - Business Systems

GitHub Activity

GitHub streak

GitHub contribution chart

Contact

Pinned Loading

  1. Rubiks-Cube Rubiks-Cube Public

    A beginner-friendly Three.js project demonstrating a rotating 3D cube with interactive controls.

    JavaScript 1

  2. UniProof UniProof Public

    Evidence-first AI university research and application guidance

    TypeScript 1

  3. codex-chatgpt-control-plus-subscription codex-chatgpt-control-plus-subscription Public

    Forked from adamallcock/codex-chatgpt-control

    Unofficial SDK for Codex agents controlling visible ChatGPT web sessions. For ChatGPT Plus subscribers with working thread recovery if browser-kernel execution timed out.

    JavaScript 1

  4. OCR-Model OCR-Model Public

    OCR Model - vision information-extraction pre-model with rotation-zone clustering.

    Python

  5. vinyl_record_store_frontend vinyl_record_store_frontend Public

    Frontend for an academic vinyl record store recommender system.

    JavaScript 1

  6. agent-skills agent-skills Public

    Skills I use for my Agents (Claude Code, Codex, Copilot)

    Python 9 2