I'm Loh Kah Meng. I build AI-powered decision support systems — tools designed to sit beside human judgment, not in front of it.
Every project on this site follows the same shape: start from real, messy data; build a governed, auditable pipeline rather than a one-off script; and end with something a human decision-maker can actually use — a dashboard, a scored list, a clear explanation — rather than a model that just sits in a notebook. That applies whether the domain is workforce analytics, healthcare, credit risk, or customer retention.
I think the most important design decision in any AI system isn't the architecture — it's where the human sits relative to the model. A system that hides its reasoning and expects to be trusted on faith is fragile the moment it's wrong. A system that shows its work, flags what it's uncertain about, and hands a clear, well-supported decision back to the person accountable for it — that's the relationship worth building toward, and it's the standard I hold every project on this site to.
That standard isn't abstract to me — I trace it back to Prof. Dhanjoo N. Ghista, a founding figure of biomedical engineering who was senior professor on the biomedical engineering program at Nanyang Technological University, Singapore while I was there. A full tribute to his work and influence is here.
I'm always glad to talk about a project in more depth, share the underlying code, or hear feedback on the write-ups here.