Singapore · bank by day, builder by night
I spend my day job keeping a bank's social media from catching fire across seven countries. At night, I teach myself to code and ship things that mostly work.
Selected work
Vibe-coded end-to-end, usually at 11pm, and somehow all still live.
A fair-value model for Singapore property: Fair PSF = baseline × (remaining lease / 99)^0.6, checked in real time against official URA/HDB transactions, plus rental-yield and social-sentiment signals — so you can sanity-check the agent's pitch before you believe it.
A monthly scorecard for the US and Singapore markets across trend, valuation, breadth, stability and the economy — with a stock screener and cross-market comparisons. Built to answer "should I be worried?" before I actually spiral.
Scrapes and scores MOE's Primary 1 balloting and vacancy data across all 182 Singapore primary schools. The odds engine is kept deliberately separate from the AI narration layer, because stressed-out parents deserve real numbers, not hallucinated ones.
A dashboard tracking news and social coverage of the Singapore banking sector — automated ingestion, LLM tagging by bank/theme/sentiment, and AI-written executive briefings, so nobody has to read everything by hand. That someone used to be me.
About
I've spent 15+ years in enterprise social media and martech at two of Singapore's biggest banks (UOB and DBS) — going from running a single Facebook page to heading a Social Media Command Centre across 7 markets, mostly through stubbornness and a refusal to let a platform migration beat me.
Somewhere along the way, clicking buttons in enterprise software all day made me want to build my own buttons. So I taught myself to code, and now I spend my evenings shipping real, public, occasionally-embarrassing-when-they-break products: property valuation models, market dashboards, a school-balloting calculator for panicking parents — held together by Python, FastAPI, a growing fondness for the Model Context Protocol, and an unreasonable amount of coffee.
I'm currently doing a Specialist Diploma in Applied Generative AI, mostly so I have official paper backing up what I was already doing for fun. Before all of this, I was an electrical engineer by degree, which explains approximately none of my career path.
For the formal version — titles, dates, the whole résumé — LinkedIn has it. This site is the unofficial one.
What I'm actually good at
15 years across a lot of different hats. Here's the honest breakdown.
I write the policies nobody reads until something goes wrong — governance frameworks, crisis response plans, escalation workflows, and the risk-monitoring keyword lists that catch trouble before it becomes a headline.
Google Trends, keyword planners and social listening tools aren't just for marketers — I use them to read market and consumer signals most people don't think to look for.
I've shot and edited Reels, ghostwritten for clients, written LinkedIn thought pieces, and made more Photoshop, Canva and Figma graphics than I'll ever admit to.
Pixels, conversion events, A/B tests, campaign measurement — comfortable running paid media across Meta, TikTok, LinkedIn and Google Ads, and actually reading what the numbers say afterward.
I train teams across markets on social strategy and crisis handling for a living. For fun, I run coffee appreciation workshops — a much lower-stakes audience.
Self-taught, Python-first, occasionally janky. I've shipped a proptech startup and the four products above, building and rebuilding my own tools along the way.
From the ML/DL fundamentals — CNNs, GANs — to today's LLMs and applied GenAI at work, I like understanding how the model actually works before I trust what it tells me.
Off duty
I mean, I'm often still staring at a screen. Just a different one.
Over-analysing pour-over ratios like they're a KPI dashboard.
Working on hitting the ball in the direction I actually intended.
DSLR in hand, mostly street shots nobody commissioned.
Slowly learning how to not get hit. Progress is... ongoing.
Get in touch
Got a project idea, feedback on one of the sites above, or a strong opinion about pour-over ratios? I'm around.