The robot that ships code

Six months after joining, Arjun was leading deployments. Every Friday evening the team would push code, then someone would manually run tests, build the Docker image, SSH into the server, pull the latest code, restart the service, and pray.

It took two hours every time. Once, Vikram forgot to run tests. A bug went live at 6pm on a Friday. It took until Sunday to fix.

"There has to be a better way," Arjun told his manager.

"There is," she said. "It's called CI/CD."

What is CI/CD?

CI — Continuous Integration: Every time someone pushes code, an automated system immediately runs all the tests. If tests fail, the code doesn't proceed. This catches bugs before they reach production.

CD — Continuous Deployment: If tests pass, the code is automatically built, packaged, and deployed to the server. No manual steps. No forgotten tests. No Friday evening rituals.

Think of CI as the quality inspector on a factory line — every product gets checked before moving forward. CD is the conveyor belt that delivers the approved product to the customer automatically.

GitHub Actions — the most popular CI/CD tool

GitHub Actions runs workflows defined in YAML files inside your repository. Every push to the main branch triggers the workflow.

```yaml

.github/workflows/deploy.yml

name: Test and Deploy

on:

push:

branches: [main]

jobs:

test:

runs-on: ubuntu-latest
steps:
  - name: Get the code
    uses: actions/checkout@v3

  - name: Setup Python
    uses: actions/setup-python@v4
    with:
      python-version: '3.11'

  - name: Install dependencies
    run: pip install -r requirements.txt

  - name: Run tests
    run: pytest tests/

deploy:

needs: test  # Only runs if tests pass
runs-on: ubuntu-latest
steps:
  - name: Deploy to server
    run: |
      ssh user@server "
        cd /app &&
        git pull &&
        docker compose up -d --build
      "

```

What happens every time someone pushes code now:

1. GitHub Actions starts automatically — no human trigger needed

2. It installs all dependencies on a fresh virtual machine

3. It runs every single test

4. If any test fails — it stops and sends an email. Nothing gets deployed.

5. If all tests pass — it SSHes into the server and deploys automatically

6. The whole thing takes 4 minutes

Arjun's Friday evenings became free. The team deployed 3-4 times a day instead of once a week. Bugs got caught in minutes, not days.

The stages of a real pipeline

```

Push code → Lint (code style check) → Unit tests →

Integration tests → Build Docker image →

Push to registry → Deploy to staging →

Run smoke tests → Deploy to production

```

Not every team does all of these. But every team should at minimum: run tests and deploy automatically.

Arjun's rule for new teams: Start simple. One workflow. Three steps: checkout → test → deploy. Add complexity only when you feel pain.