Problem
Finding remote backend/mobile roles is noisy: offers are spread across many boards, repeated across sources, and rarely filtered to your preferences.
What I built
A lightweight job aggregator + matcher:
- Ingests job postings from multiple RSS feeds
- Stores them in PostgreSQL
- Lets you create profiles (keywords + excluded terms)
- Returns ranked matches with a simple scoring model
Screenshots
Architecture
- FastAPI API (
/ingest,/jobs,/profiles,/matches) - Postgres in Docker for local persistence
- SQLAlchemy for data access
- Minimal UI served by the backend for quick visualization
Challenges & tradeoffs
- RSS feeds vary a lot in structure, so normalization (company/location) is best-effort.
- Kept the first version simple: scoring is keyword-based (fast + explainable), not ML.
Results / Impact
- One command to populate the DB (
POST /ingest) - Profiles instantly surface the most relevant jobs without manual searching
What I’d improve next
- CRUD for RSS sources
- Better parsing/normalization (company, location, contract type)
- More filters (timezone, English-only, remote-only, salary)
- Scheduled ingest + notifications (email/telegram)
- Full-text search and saved searches