Health & Safety System Prototype

Built for a real client — PricewaterhouseCoopers — through UBC's CPSC 319 software engineering project course. The ask was a Health and Safety incident system frontline teams could actually use to log, track, and understand safety incidents, delivered by an 8-person student team on a hard course deadline.

Full-stack on Google Cloud: Next.js on the front end, Firestore as the NoSQL data layer, and Cloud Run for serverless hosting. Serverless and NoSQL were the right call for a student team working against a fixed deadline — infrastructure we didn't have to manage ourselves, and a data model flexible enough to change as we learned more about how incidents actually get reported.

The feature that pushed the prototype past a standard CRUD app was a generative AI chatbot built on Vertex AI, layered over the incident data for trend analysis in plain language instead of manual filtering. It cut query response time by about 25% — the kind of improvement that matters most when someone's trying to catch a pattern before it becomes a bigger problem.