Projects

Things I built while learning, exploring, and trying to understand the world better.

01

Seeing

Exploring how AI can perceive, recognize, and connect information with real-world context.

Innovation Concept · AI Application

AI Image Search

An AI-assisted innovation concept exploring how image recognition can simplify the process of understanding unfamiliar buildings and rental environments.

The idea originated from research into Melbourne's rental challenges, especially for newcomers and international students who often struggle to recognize buildings, compare rental information, and understand surrounding environments.

I explored how computer vision, contextual search, and integrated information systems could create a more intuitive rental discovery experience through AI-powered image search.

User ResearchAI ApplicationInnovation Concept

02

Understanding

Using structured logic, data, and AI-assisted reasoning to support clearer understanding and decision-making.

AI Learning System · Personal Project

PTE AI Learning Advisor

A prototype system designed for PTE learners to analyse score trends across listening, reading, speaking, and writing, identify weak areas, and generate personalised study suggestions.

I designed the score data structure, processed simulated score data with Python and Pandas, and created Tableau visualisations for score trends, skill comparison, and historical performance analysis.

The system connects AI API with prompt design to generate personalised learning advice based on user performance, while also planning future metrics such as user activity, progress rate, and target achievement rate.

AI APITableauLearning Analytics

Decision Support · Internship Project

Intelligent Scoring Decision Support System

A decision support system developed during my internship at 2Cloud Technology, focused on structured scoring logic, user input flow, and result explanation.

I worked on multi-step data collection, rule-based scoring logic, and weak-point analysis to help users better understand their score structure and possible improvement directions.

I also used Tableau to improve the result presentation through visual elements such as radar charts and progress indicators, making complex scoring results easier to interpret.

Decision SupportRule-based LogicTableau

Machine Learning · Kaggle Competition

Salary Range Classification

A Kaggle multi-class classification project based on 100,000+ recruitment records, aiming to predict salary ranges from job-related and demographic features.

I worked on missing value handling, outlier analysis, feature distribution visualisation, categorical feature processing, and cross-feature construction such as job-industry combinations.

I experimented with LightGBM, XGBoost, logistic regression, and stacking-based model fusion, achieving 1st place on the public leaderboard and 22nd on the private leaderboard among 249 teams.

ClassificationFeature EngineeringLightGBM

03

Building

Turning ideas, systems, and personal reflections into interactive digital experiences.

Design System · Personal Space

Personal Website

A personal digital space exploring AI, growth, uncertainty, and creative identity through a minimal and reflective design system.

This website is designed less as a traditional portfolio and more as a personal space for ideas, projects, and the development of my own creative IP.

Next.jsDesign SystemPersonal IP