B.Eng. Engineering Product Development · SUTD · Class of 2029
I am an engineering undergraduate in Singapore and an ASEAN Undergraduate Scholar. I build things that have to survive contact with a real user — a deployed trading tool, a doorbell that never touches mains power, an autonomous robot, a marketplace designed on the island it serves. Every one of them is written up here with the part that went wrong still in it. Also, depending on the hour, a photographer and a perfect-pitch pianist.
Selected work
Each one is written up in full on the projects page — what the problem actually was, what I got wrong first, and what survived contact with a real user.
01 · Deployed
A trading journal and fee calculator for Indonesian retail investors on the IDX. My mother keeps her trades across several notebooks; this is the tool I wished existed for her.
48tests, each one written to prove it refuses bad input
Where it went wrongOf three defects my AI teammate reported, two were not real. I only knew because I read the source.
See the product page02 · Energy systems
Sized by photovoltaic and battery calculation against the worst week it would ever see rather than the average one.
0watts drawn from mains power, ever
Where it went wrongMy first enclosure sealed the panel in flush. It muffled its own speaker and cooked the electronics.
Read the write-up03 · Global immersion
A marketplace for Balinese traditional performers, designed in Bali with Institut Teknologi Sepuluh Nopember.
30–40%of every booking fee lost to middlemen
Where it went wrongI was designing for a country I am from, and found I had never actually examined it.
Read the write-up04 · Product design
A carriage-occupancy display so MRT passengers stop packing one cabin while the next one runs half empty.
0 → 100%detection, once we changed the input
Where it went wrongThe first test detected nothing at all. The instinct was to blame the model; the fault was what we showed it.
Read the write-up05 · Deep tech
Two days to turn a live quantum-computing research project into something a fund could actually back.
TRL 3 → 9the span our proposed roadmap had to cover, across three hardware families
Where it went wrongTwo days is not enough to understand quantum error characterisation. We said so, to the researcher who wrote it.
Read the write-up06 · Embedded
A medication reminder for elderly users with no smartphone and no dependable connection — ESP32, a web app, and a serial bridge between them.
122,000Singaporeans over 65 living alone by 2030
Where it went wrongThe bridge died inside the device's memory budget, on stage, mid-demonstration. We diagnosed it live.
Read the write-upOff the bench
I have never believed the engineering and the music are separate activities. Both are the same instinct — take something structured, find where it bends, and rebuild it so it sounds right.
In progress
This site replaces my old Google Site. Everything in the navigation above is live; the rest arrives chapter by chapter.
