All Projects

AI pose detection module that automates image capture to find the perfect suit size using Tensorflow.js and PoseNet

Mission-critical digital signage platform serving 250,000+ screens globally with 24/7 uptime

Scalable vaccine distribution system for Thai hospitals handling millions of concurrent users

Premium steakhouse delivery platform for UP A Notch, featuring seamless online ordering, real-time order tracking, and LINE integration. Built as a lightning-fast PWA with a modern, scalable cloud backend. Elevated the restaurant's digital presence and customer experience.

Hackathon 2nd place winner (2017). An accessibility app that voices out the medication in front of the camera, empowering blind and visually impaired users to independently identify their medicine. Utilizes real-time image recognition and text-to-speech to provide instant audio feedback.

1st place winner (2024). A prototype app for Sansri, a leading building construction company, to solve pain points in worker management. The system enables real-time on-site worker tracking and visualizes building progress, streamlining construction operations and improving transparency.

A comprehensive benchmark of machine learning algorithms for predicting Polycystic Ovary Syndrome (PCOS). Compared Logistic Regression, Random Forest, and Decision Tree models to identify the most effective approach for PCOS diagnosis using real-world clinical datasets.

An interactive Tic Tac Toe game powered by Deep Q-Learning, built with React, TypeScript, TailwindCSS, and TensorFlow.js. The AI starts with zero knowledge and learns in real-time as users play, improving its strategy with every move. This project demonstrates reinforcement learning concepts in a fun, visual, and hands-on way directly in the browser.

A high-performance, real-time financial analytics platform for professional traders and analysts, developed at LSEG. Workspace delivers millions of data points per minute with a seamless desktop experience, leveraging Electron, TypeScript, Redux, and IndexedDB for robust state management and offline capabilities. My contributions focused on optimizing data pipelines, reducing load times, and enhancing user experience for mission-critical workflows.

A comprehensive web platform for the Royal College of Surgeons of Thailand (RCST) to facilitate surgeon registration, examination scheduling, and license management. The system enables doctors to sign up, schedule their licensing examinations, and receive results. Administrators can manage users, educational resources, examination dates and venues, and publish pass/fail results. Built with a modern stack to ensure reliability, scalability, and security for critical medical certification workflows.

Semifinalist (top 10 out of 2500+ teams globally) in the IBM Call For Code Hackathon. Invented and developed a pre & post worldwide disaster solution platform to aid communities before, during, and after disasters. The platform leverages IBM Cloud, Watson NLU, Android, Java, Docker, and MongoDB to provide real-time information, resource coordination, and communication for affected populations and responders.

Designed and deployed a Jupyter-powered digital learning environment at Cal State Fullerton, empowering over 500 undergraduate students and 80 faculty with hands-on access to Python, AI, and data science tools. The platform cultivates data literacy, AI fluency, and computational proficiency by integrating real-world datasets and practical assignments into STEM and interdisciplinary courses. Project outcomes include measurable gains in student engagement, AI and data skills, and increased adoption across multiple departments through course modules and faculty professional development.

Developed a fully automated PDF accessibility remediation platform for Cal State Fullerton, empowering students, faculty, and staff with accessible digital content in compliance with ADA and Section 508. This tool leverages AWS Bedrock, Lambda, Cognito, and a modern React frontend to deliver scalable, AI-powered remediation of PDF documents for people with disabilities. Users securely login and can submit PDF files, which are processed to meet regulatory accessibility standards (such as readable text, proper tagging, and alternative text), ensuring equitable access to essential information and resources campus-wide.