Data Analytics & Computer Vision Engineer
Information Engineering & Media student at NTU, building at the intersection of intelligent systems and human-centred design. I turn messy data and complex problems into things that actually work.
I'm a second-year Information Engineering & Media student at Nanyang Technological University (NTU), currently on exchange at NUS focusing on Innovation, Product Design & Software Engineering. I'm on track for First Class Honours, with academic interests spanning both technical engineering and the design side of media and product.
My work sits at the crossroads of AI systems and real-world impact — whether that's building computer vision pipelines for robotics, engineering data-driven marketing strategies, or designing assistive technology for people with disabilities.
Outside of tech, I lead media and partnerships for Indonesia Cultural Night and contribute to the EEE Club's design and communications. I believe the best solutions are built by people who understand both the system and the human on the other side of it.
A Java-based inventory management system built as part of a 5-person team project, helping small businesses track stock levels, manage items, and stay on top of inventory through a fast, command-line-driven workflow with automatic low-stock alerts and category-based organisation.
Engineered a computer vision pipeline using MTCNN and FaceNet for face detection and feature embedding, deployed on a robot platform with OpenCV optimised for real-time performance.
Full-cycle data science project on Kaggle — from EDA and feature engineering to model selection. Benchmarked Linear Regression, XGBoost, HistGradientBoosting, and LightGBM with 5-fold cross-validation and hyperparameter tuning; LightGBM came out on top with an R² of 0.92.
A dynamic web application built with Java Servlets and Apache Tomcat, featuring robust back-end inventory management, secure user transactions, and modular design patterns.
Designed and prototyped an immersive 3D horror environment in Unity, with custom C# scripting for interactive elements. Players must search a derelict house for a key, avoid a hidden threat, and use objects like a lantern and gun to survive — built from scene composition and lighting to particle effects and player interaction logic.
A capstone analytics project (simulated client, Watershed Properties) evaluating short-term rental conversion potential across 244 properties. Built a linear regression model to predict occupancy from rent percentile, then used Excel Solver to optimize rent and identify 41 profitable properties worth $1.37M in year, and presented via an interactive Tableau dashboard.
A collection of four full-cycle data analytics projects spanning classification, time series forecasting, NLP, and clustering: fraud detection on 285K credit card transactions (Random Forest vs Logistic Regression), Walmart weekly sales forecasting (Holt-Winters vs seasonal-naive baseline), IMDb review sentiment classification (TF-IDF + Logistic Regression, 89.2% accuracy), and mall customer segmentation (K-Means). Each covers EDA, feature engineering, and model evaluation, backed by interactive dashboards in Power BI and Tableau.
Open to internships, collaborations, and interesting conversations — especially at the intersection of AI, data, and real-world impact.