A pipeline that scrapes college notice boards, matches new notices against a lattice-based scope hierarchy (institution → faculty → class → section → individual), and pushes alerts the moment something relevant is published — no more refreshing a notice board by hand. Delivery runs on Firebase Cloud Messaging, auth is handled by microservices I wrote in Go, and I've been prototyping a React Native front end so it works as a real mobile app.

A URL shortener backend written in Go — takes a long URL, generates a short code, and redirects on lookup. Deployed and running live, built to keep the core service small, fast, and dependency-light.

Local government notices in Nepal mostly live on ward and municipality Facebook pages, where feed ranking buries them. This flips that: residents register their ward once, and new notices get pushed straight to them over SMS or Telegram — no algorithm in the way. Designed to scale across all 753 municipalities without changing the delivery model.

Hands-on practice attacking the same kind of systems I build: OAuth flows (PKCE bypass, open redirect, CSRF, scope escalation), JWT validation edge cases, and static analysis on Go services. Working through PortSwigger Web Academy labs and live scope on HackerOne.