A systems analyst turned data engineer tried to schedule an ETL pipeline built in Google Colab. The script depended on a hardcoded Colab‑mounted path, which broke outside that environment. By replacing the path with an environment variable, the author made the pipeline portable and runnable as a plain Python script anywhere.

Once portable, the author tested the script locally, confirming it created a SQLite database and logged completion. With a working standalone script, the next step was choosing an orchestration tool. GitHub Actions emerged as a lightweight option: it runs on GitHub's servers, uses a YAML workflow, and requires no extra infrastructure, fitting the project’s scale.

Deploying the workflow, the author set a cron trigger for 9 AM UTC daily and added a manual dispatch button for testing. The first run finished in 27 seconds, loading data into SQLite on a GitHub runner. Updating deprecated actions fixed a warning, and the pipeline now runs cleanly every day without manual intervention.

The exercise highlighted that true scheduling begins with portability. A pipeline tied to a single notebook platform cannot scale. By abstracting file paths and leveraging GitHub Actions, the author achieved repeatable, infrastructure‑free runs. This approach demonstrates how simple environment tweaks and cloud‑hosted workflows can transform a hobby notebook into a production‑ready data pipeline for data engineering teams looking to cut costs and time effort.