Newsletter Radar is a self-initiated SaaS project to help people regain control of their Gmail inbox. It automatically detects newsletters and generates AI-powered, personalized digests, built to demonstrate the ability to design, build, and ship a working product end-to-end, beyond client-facing analytics engineering work.
What I did
Built a secure Gmail OAuth2 connection (no shared passwords)
Designed intelligent newsletter detection beyond sender matching: technical headers (e.g. List-Unsubscribe), unsubscribe links, content patterns, and known sender domains
Built an AI digest pipeline (Claude API): interest-based filtering, key-point extraction, and short summaries (5 to 10 points) linking back to the source email
Added inbox organization features: automatic "Newsletters" labeling, mark-as-read option, and a "deep clean" scan for processing years of inbox history
Outcome
A working, deployed product that turns inbox overload into a quick, readable digest, showing the same automation, API integration, and pipeline thinking used in client analytics engineering work, applied to a self-built product.
A self-built SaaS app that connects to Gmail via OAuth2, auto-detects newsletters, and uses AI (Claude) to generate personalized digests, a demonstration of end-to-end product and automation skills.