I architect AI systems
for the real world.
CS student at PolyU. I design secure, end-to-end AI pipelines from edge hardware to cloud infrastructure. Built a waste detection system for Cathay Pacific using RFDETR Nano + FiLM on-device. Founded CUBIK Studio.
See What I've Built
AI Tutor Platform
RAG pipeline with Ollama grounding LLM responses in course materials for COMP4431
Authored the full system architecture for a web-based AI Tutor platform for COMP4431. Engineered a RAG pipeline utilizing Ollama to strictly ground LLM responses in course materials, minimizing hallucinations and ensuring high accuracy.
Cathay Pacific Hackathon 2024
2nd Runner Up — Designed waste sorting trolley system architecture for Cathay Pacific
Designed the waste sorting trolley system architecture using RFDETR Nano and FiLM combined with MobileNet for Cathay Pacific. Implemented the full detection pipeline for the hackathon demo.
Smart Home IoT System
ESP32 edge computing with MQTT, Node-RED, InfluxDB, Grafana, and Python ML predictions on Docker.
Interactive Polling System
Node.js & AWS DynamoDB web platform for real-time polls with JWT auth, live updates, and AWS QuickSight analytics.
RAICOM 2024 — 1st Prize
Developed Computer Vision model with YOLOv8 for the robot treasure-hunting autonomous task competition.
ABU Robocon 2023 & 2024
Controlled shooter pitch/yaw with Arduino & STM32, developed CAN-Bus library, supervised swerve & CyberGear systems on ROS Linux.
Year 4 BSc (Hons) Computer Science student at The Hong Kong Polytechnic University with a minor in Information Security. Founder & CTO of CUBIK Studio Limited, where I design and implement AI-powered systems for enterprise clients. Passionate about building robust, end-to-end solutions across AI, full-stack development, and embedded systems.