Skip to main content

About

Loh John Ray. Year 3 Engineering Science at Ngee Ann Polytechnic, working somewhere between deep-tech research and human-centred engineering.

Who I am today is an amorphous plurality of beliefs, traits and values that is constantly updating. I think these experiences inculcated my most prominent traits in science and engineering, but it would be myopic to claim I have anywhere near a complete picture of who I actually am, or who I will become. What follows is an attempt to draw the connections, crudely.

I chose Engineering Science for a slightly unserious reason — forty people, two classes, goated lecturers — and stayed for a serious one. I remember my first semester in 38PROG. Twenty or thirty people in the room and it was dead silent; all I could hear was my own keyboard. That was when I realised ES was no joke.

There is a communal room, 08-03-05, where ES students hang out, study, and put events together. Most of what I know about working with people I learned in rooms like that one, and then re-learned the hard way in a committee.

My research sits in CO₂ adsorption — packed-bed columns, polymer sorbents, breakthrough curves for Direct Air Capture. My builds sit elsewhere: multimodal AI for misinformation, a voice-directed surgical co-pilot, an open-source blood sampler you can print tonight. Both, at once, all the way down.

History

  1. Why Engineering Science? Forty people, two classes, goated lecturers. I walked in wanting to make friends and found out half the cohort wanted the same thing. It took one hello.

  2. Description, unedited: 'beginner programmer; want to make a working integral calculator.' Still unfinished. Still pushed to this year.

  3. First time abroad in ten years. Facilitated nightly reflection, badly at first. Learned that things are rarely as they appear outwardly.

  4. Gachapon machines, RC cars, terrariums, clay jewellery. Learned to stop solving other people's problems for them.

  5. A social hackathon, 18 hours teaching digital literacy to isolated seniors, and Saturday mathematics with children from lower-income families.

  6. Three ES students and one aerospace student, cobbled together after a school-wide email. We thought we had lost, went home early, and missed the prize ceremony.

  7. Retractable tactile paving. Junior Design Innovator (Advanced), then a S$50k innovation fund shortlist. We did not get the money.

  8. 700+ Leos, 50+ school-approved event proposals, monthly minutes to Lions International Singapore, and a lot of 5.45am Saturdays.

  9. AI-generated classroom posters pushed to an ESP32 OLED. First full-stack build. The hardware half never quite landed.

  10. SENTINEL, built in 24 hours and then extended: multimodal detection on Cloud Run, Telegram bot, SEA-LION GUARD for the languages Singapore actually speaks.

  11. Two weeks back to back. One session until 4am. Learned that the brief is a conversation, and that listening is the hard part, not the engineering.

  12. A sub-S$10, fully 3D-printable blood sampler for disaster and low-resource settings. Released under CC-BY-4.0.

  13. Packed-bed breakthrough with polymer sorbents, supervised by Prof. Erik Birgersson (NUS) and Dr. Prapatsorn Borisut (SUTD). Scope revised in April; three validation gates before any headline result.

  14. Mentors from Schneider Electric, Constellar and LinkedIn, and a three-month team project on school energy management. Where I decided to create luck through volume.

  15. A voice-directed surgical co-pilot, a handwriting-to-LaTeX pipeline, and a digital employee for heartland merchants. Three hackathon-shaped months.

  16. Applied on behalf of the school. No local rep to call, ten people to recruit, sent the email anyway.

Things I am still bad at

Communicating an idea under pressure. The need to compress everything I know into a few sentences is, at times, palpable, and it is the single skill I am most deliberately working on.

Giving feedback that is useful rather than nice. Steering a reflection without steering it into my own conclusion. Waiting ninety seconds before offering a fix.

I applied to DSTA once and reached the interview stage — I think the portfolio and a 3.95 GPA carried me there. I talked about compliant mechanisms and computational design with neural networks, and I was reading Learning Deep Learning at the time. I did not get it. I am applying again with a lower GPA, a much stronger portfolio, and a clearer reason: I want to start doing research now, not later.

Elsewhere