Selected projects
I move between machine-learning research and product engineering: framing the problem, building the system, testing the difficult parts, and giving the result an interface people can actually use.
Portfolio
No imagined roadmap and no placeholder case studies — these are systems currently in the workspace, each at an honest stage of development.
Trust-aware grape-leaf disease assessment that pairs classification with calibration, visual explanation, and unfamiliar-disease detection.
An offline decision-support workflow for nurses that turns intake information into a structured triage recommendation without sending patient data to the cloud.
A virtual-makeup pipeline designed to enhance a portrait without erasing texture, changing facial geometry, or shifting skin tone.
A local-first analytics platform for one creator account, using official TikTok OAuth and APIs while keeping private Studio metrics clearly separated from public data.
A guided coffee-shop ordering experience that recommends drinks and food from mood, hunger, dietary preferences, and context.
A hand-built personal world that carries research, photography, writing, and identity without flattening them into a generic template.
Across the work
I turn broad questions into experiments, baselines, failure tests, and evidence someone else can inspect.
I work across data, models, APIs, databases, interfaces, tests, and deployment instead of stopping at the notebook.
I design around the person making the decision: what they need to see, control, understand, and trust.
Privacy, fairness, accessibility, and honest status labels are part of the build from the beginning.
Work together
I am open to research collaborations, applied-ML roles, and thoughtful product work.
Get in touch