Job
Tracker

An open-source kanban board to track job applications: each job posting is a card that moves through fixed stages of the hiring process, with a per-card event timeline and automatic history. Built to organize my own job search.

View on GitHub ← All Projects

// About the project

What was built

Job Tracker is a kanban board to track job applications: each job posting is a card that moves through fixed stages of the hiring process (Interested → Applied → In Process → Offer → Closed), with a per-card event timeline and automatic history on every stage change.

I built the project to organize my own job search, and decided to keep the code open source as a sample of how I structure a full-stack project from scratch: custom cookie-session authentication, Google login, PostgreSQL persistence with raw SQL and versioned migrations, and a real integration test suite against a Postgres instance in Docker.

The repository runs a full CI pipeline on every pull request: Jest tests, ESLint and Prettier linting, Secretlint secret scanning, and continuous deployment on Vercel.

// Challenges

What made the project difficult

// Modeling

Fixed stages and events

Modeling the domain so every stage change automatically generates a timeline event, without duplicating logic between the card and its history.

// Auth

Custom sessions + Google

Implementing cookie-based sessions with bcrypt and a pepper from scratch, and integrating Google OAuth login, keeping both auth paths consistent.

// Testing

TDD end to end

Every endpoint is born with a success and an error test before the production code, running against a real Next.js server and a real Postgres via Docker.

// CI

Full pipeline

Setting up Jest, ESLint, Prettier, Secretlint, and Vercel preview deploys to run automatically on every pull request, without adding friction to my own commit workflow.

// Tools

Technical stack

Next.js

PostgreSQL

Jest

Docker

Google OAuth

GitHub Actions

Vercel

// Outcome

What I learned

Building your own tool: using a personal project to solve a real day-to-day problem makes every product decision more concrete.

CI from the first commit: setting up linting, tests, and secret scanning early prevents problems from piling up as the project grows.

Open source raises the quality bar: knowing the repository is public reinforces discipline around tests, commits, and documentation.

TDD reduces rework: writing the test before the endpoint makes the exact shape of the response clear before implementing any logic.

JobTracker

Job
Tracker