Making software reliable enough to trust with real decisions. Now researching whether AI can do the same.
Software engineer with production experience building data-intensive systems across regulated domains — including government service eligibility platforms, health commerce platforms, compliance-governed payment and identity verification systems, and biosecurity data pipelines. My work has consistently involved a specific engineering challenge: translating complex domain requirements, regulatory obligations, data integrity constraints, and compliance rules into reliable, testable, and auditable software logic. Alongside this, I contributed to a published team research project applying machine learning to a health data problem, developing a prediction pipeline for maternal morbidity outcomes using the Bangladesh Demographic and Health Survey dataset. That project covered model selection, API deployment, and clinical output visualisation, making visible a gap that production engineering does not close — the methodological rigour needed to evaluate whether a model is actually trustworthy, not just whether it runs. I am now pursuing postgraduate research to develop that depth formally, at the intersection of applied AI, health data systems, and regulated biological domains.
The NSW eID transition is landing as a real compliance burden — not just for commercial producers, but for smallholders too. From backyard PIC thresholds to strict-liability penalties and the looming 2027 all-digital cutoff, here are four realities most people didn’t see coming.
Team contribution addressing UN SDG Goal 3. Built a Random Forest model achieving ~95% accuracy for fetal health prediction, deployed as a REST API via FastAPI and Heroku, with a React/TypeScript data visualisation dashboard.