About / Story

I started in bio-photonics, building and validating ML models for intraoperative tissue classification using Raman spectroscopy. That work was published in Lasers in Surgery and Medicine, received Editor's Choice distinction, and was presented at Harvard's National Collegiate Research Conference. A recurring lesson from that work was interpretability: a model that a clinician cannot understand is a model a clinician cannot trust.

That theme runs through my current work. My capstone project applies ML to gut microbiome data for colorectal cancer detection, with cross-cohort validation across Malaysian, Sri Lankan, and Saudi Arabian populations. It is a deliberate stress test of whether clinical AI systems actually generalize beyond the datasets they were trained on. Beyond the lab, I've built production ML pipelines at scale, shipped software into clinical environments, and spent time in the field observing what happens when diagnostic infrastructure is absent entirely.

I'm also shaped by a fairly global life across Thailand, India, the UAE, and the U.S.; outside work, I watch a lot of football, support Manchester United, and can never really say no to gelato.