Capstone Lead Developer · 2025
A health platform built to take the friction out of logging food and habits by hand, using image-based calorie counting, lifestyle prediction and generated summaries.
Manual food and health logging is the reason people stop. The capstone brief was to remove it, and to do that with three separate streams of work: an AI team building the models, a data team collecting and presenting data, and the developers putting it in front of a user.
Five people, and I led the development side with one other developer. The part I owned was deciding how the backend and frontend work was split, and then integrating the output of the other two groups into the actual product. It was not a web app with a feature bolted on, it was three streams of work that had to meet in one place.
Delivered as the capstone of the Coding Camp powered by DBS Foundation (Dicoding) full-stack cohort, with the ML pipelines, backend APIs and interface meeting in one working product.
How to communicate with people who are not developers, efficiently and without friction, which mostly meant getting specific about interfaces early instead of discovering mismatches late. This also ran from just before finals straight through them, so I learned to hold a project steady while the rest of my schedule was at its worst.