
Open-source, non-profit student platform for NTHU centralizing courses, grades, bus schedules, and announcements. Grew to 10,000+ users (8,800 peak MAU, ~60% of student body). Built with React, PostgreSQL, Docker, Cloud VMs, and Cloudflare Workers.
Architected core perception (Jetson Nano with ROS2, OpenCV for pathfinding/object recognition) and control (ESP32 firmware with PID for precise movement). Achieved 100% reliability in obstacle course with zero errors under competition conditions.
Led perception and inference software stack for AI food inventory system using Gemini API, TensorFlow, React, 3D printing, Arduino, and Raspberry Pi. Developed closed-loop prototype with proactive alerts and recipe generation. Highlighted for 'Best-in-Class' Innovation.
An IoT System that connects the Makerspace lights in our high school club to be controlled by Google Assistant.
End-to-end digital design flow for a 3-bit histogram adder chip using 0.18 µm CMOS node. Completed circuit design, HSpice/Laker/Calibre simulations (DRC/LVS) with full verification. Top 10% of class.
Low-cost table tracking system using fiducial markers (printed on receipts) and central PixArt imaging sensor. Team-led prototype for fast-food chains achieving high accuracy as alternative to expensive Bluetooth solutions.
A system that manages our club's management duties like students, attendance, finance, inventory and so much more.
A semiconductor equipment control protocol that is used to network manufacturing equipments. This is a host that manages the protocol.
Designed and deployed 56 synchronized clocks with temperature/humidity sensors using ESP32, Firebase, MQTT. Built dashboard for live monitoring and OTA updates, solving time-drift issues - still deployed at Chung Ling Private High School.
A mentorship site for economic students that aims to become a social platform
A word clock that displays the time in words.
A levitator that is able to levitate objects up in mid air using ultrasonic sound.
A smart bin that keeps track of the weight of the bin and the amount of waste in it.
A bus arrival system that displays the next bus arrival time and the bus number.
An 8×8×8 LED Cube controlled with an Arduino and 74HC595Ns.
An IoT System that connects the Makerspace lights in our high school club to be controlled by Google Assistant.
Twist App is a digital receipt system that aimed to reduce the amount of paper used by printing receipts.
A livestock attendance system that is directionally aware attendance system using Long Range RFIDs.
An object avoidance system for a prototype drone during my internship at Vitrox
Award-winning safety device with custom LED matrix panel for traffic police. Designed hardware with surface-mount LEDs, Arduino Pro Mini controller, etched circuit boards, and firmware for multi-color signage display. Addressed safety issue of ambiguous hand gestures.
A directional aware attendance system using Long Range RFIDs that are embedded within school uniforms.
Participant
1st Runner Up
1st Runner Up
Overall Champion
Oct 2023 - Present
Hsinchu, Taiwan
Architected full-stack system from scratch using React, PostgreSQL, Docker, Cloud VMs, and Cloudflare Workers, centralizing courses, grades, bus schedules, and announcements for 10,000+ users (8,800 peak MAU, ~60% of student body). Recruited and managed 20+ contributors, overseeing product roadmap, technical strategy, sustainability, and succession planning. Secured fiscal sponsorship from local tech firms and advanced official talks for integration into NTHU's systems. Incorporated ongoing freelance full-stack development in 2025.
Jun 2024 - Aug 2024
Taipei, Taiwan
Led perception and inference software stack for Project 'Fresh'—AI food inventory system combating waste by tracking refrigerated items and expiry dates using Gemini API, TensorFlow, React, 3D printing, Arduino, and Raspberry Pi. Developed closed-loop prototype with proactive alerts (e.g., 'Your milk expires tomorrow') and recipe generation based on soon-to-expire items. Highlighted for 'Best-in-Class' Innovation for viable ambient AI solution.
Jun 2024 - Aug 2024
Hsinchu, Taiwan
Implemented 4 deep learning models with TensorFlow and PyTorch for EEG emotion classification, achieving 17% increased accuracy with 60% feature reduction. Collaborated on dataset analysis and model optimization for biofeedback applications.
Jun 2023 - Aug 2023
Penang, Malaysia
Process engineering internship in semiconductor manufacturing.
Jun 2022 - Aug 2022
Penang, Malaysia
Gained experience in semiconductors and radio frequency (RF) engineering.
Jan 2022 - Jun 2022
Penang, Malaysia
Implemented Semiconductor Manufacturing equipment control protocol (SECS/GEM) in 10+ factories in Penang. Developed AI Learning Kit for training Honeywell management on AI integration in production lines.
Jun 2021 - Oct 2022
Remote
Developed frontend applications using Firebase, React.js, Node.js, and Next.js.
For the most up-to-date information, please visit my LinkedIn profile at linkedin.com/in/imjustchew.