Code & Learning

Knowing every student is accounted for

With my team, I developed the Smart School Bus System in response to the risk of students being left unattended on school buses. Our research found gaps in systems that still relied on manual attendance and limited communication between schools and families. We wanted to make the journey easier to monitor—and give parents clearer information about their children’s transport.

Connecting the bus, the school, and home

Inspired by automated checkpoints at international airports, we built the system around an ESP32 microcontroller, a facial-recognition camera, and a GPS module. When a student boards or leaves, facial recognition combined with a student ID card verifies their identity and automatically records the event. The information is sent to parents and school administrators through a mobile or desktop application.
GPS tracking also makes the bus’s location visible in real time. The aim is to give families and schools a shared, timely view of the journey.

Thinking beyond a single journey

We designed the route-planning component to combine shortest-path algorithms such as Dijkstra and A* with real-time data and AI. It can recommend routes intended to reduce travel time and improve transport efficiency.
For me, the project connects computer science with urban planning at an everyday scale: how a student gets to school, how information moves between people, and how a transport system can earn their trust.

Making automation accountable

At IMCI Software, I’m designing a subscription-tracking application in TypeScript, using Prisma and PostgreSQL. I started with requirements and the software development process: understanding what needs tracking before deciding what to automate.
I’m prototyping AI workflows with the Vercel AI SDK to detect and reconcile updates across more than ten AI API subscriptions. Alongside tool-calling, I’m exploring validation rules, human approval, and audit logs—and defining how to measure accuracy and the amount of manual work saved.

Testing models against real data

I built an XGBoost model using more than 10,000 race records and an LSTM time-series model using historical data for 50 VNIndex stocks. Feature engineering and 20+ tuning experiments let me work on the less visible side of machine learning: generalization and overfitting.
This built on the Machine Learning Specialization I completed through DeepLearning.AI and Coursera in 2024, covering supervised and unsupervised learning, recommender systems, and reinforcement learning.

Explaining an invention precisely

I drafted two utility patent applications for EEG-based focus-sensing earbuds using AI signal processing. I reviewed 15+ existing patents and prepared technical specifications and diagrams. The work asked me to be precise about what an idea does, how it works, and what distinguishes it from earlier inventions.
In 2026, I received a Bronze Award in UC Berkeley’s CALICO Informatics Contest and placed in the top ten of the Open Division at the RPI Programming Competition.

Giving an idea something you can hold

I founded the Hands-on Engineering Club and developed activities for Grades 7–11, from flight mechanics and structures to rapid prototyping. I mentored students in Fusion 360, 3D printing, and Arduino. With the club, I designed and built a Plinko-inspired pull-tab collection mechanism supporting the Tondo community in the Philippines, engaging more than 60 daily visitors.
In Hanoi, I founded and led a ten-member electronics outreach club. We produced Arduino tutorials and ran kit-based workshops for students with different levels of experience. I worked on the lessons, demonstrations, equipment, and feedback that helped turn an unfamiliar subject into something students could try for themselves.

Making room for other learners

At Make The Change, I helped educators develop and deliver lessons in creativity, digital literacy, and design thinking. As Head of Event Planning on the STEM Council, I led planning for the STEM Fair and STEM Tank, helped run engineering and web-development workshops, and expanded participation through events such as Pi Day.

First Place, SJI STEM Fair, 2025.

Working through a problem—and explaining it

I study Higher Level Math AA in the IB and earned a 5 in AP Calculus BC. In 2025, I received a Bronze Honour in the International Youth Math Challenge.
As a Schoolhouse.world tutor, I worked with more than 20 students through individual and small-group SAT Math and Calculus sessions. I prepared lesson plans and practice systems to help students build accuracy and speed. Explaining a solution to someone else is also a way of finding the gaps in my own understanding.

Learning across disciplines

I attend SJI International on a full-ride scholarship. My Higher Level subjects are Math AA, Physics, and Economics; at Standard Level, I study Computer Science, Vietnamese A Literature, and English A Language & Literature.
Together, these subjects let me move between quantitative models, technology, and the ways people make decisions and communicate. Before SJI International, I attended The Olympia Schools in Hanoi on a scholarship covering 90% of tuition.