Problem
Application tracking tends to fragment across inbox threads, spreadsheets, and manual updates. I wanted a system that could ingest real job-related email activity and turn it into a structured pipeline instead of a drifting list.
Product + backend system
AI Job Application Email Tracker
Tracked 170+ applications through Outlook ingestion, stage classification, and a FastAPI-backed dashboard built to reduce drift in the hiring pipeline.
PythonFastAPINext.jsMicrosoft Graph APISQLAlchemySQLite
What I Built
I built a FastAPI-based backend that connects to Outlook through OAuth and Microsoft Graph, classifies emails by application stage, stores structured records, and feeds a dashboard designed for quick stage updates, manual entry, and backlog processing.
Key Decisions
- Kept the API surface explicit with dedicated sync, backlog, auth, and CRUD endpoints so ingestion and operator workflows stayed debuggable.
- Designed the UI around stage management first, because the real value is keeping the application board current with minimal friction.
- Combined automated email ingestion with manual entry so the tracker remained useful even when email parsing missed or delayed a signal.
Outcome / Value
- Created a working product surface with connected-email status, stage counts, search, filters, and a board view for active applications.
- Handled noisy real-world inbox data by structuring email ingestion into a reviewable classification workflow rather than a black-box sync.
- Produced a system that feels like a personal operations tool rather than a simple CRUD tracker.
Screenshots / Artifacts
Visual proof tied to the actual system.
