Introduction
In my last blog, I shared my first week experience as an MLH Fellow. This week, things got real as we officially kicked off our project!
I’m working on Apache Airflow alongside two fellow podmates. Our pod leader set up a meeting with the three of us and two Apache Airflow maintainers, who will be our mentors for the next three months. We also joined the Airflow Slack community.
A separate Slack channel was created just for us , it includes the three of us, our pod leader, and the two Airflow maintainers who will be guiding us throughout the program. This is what makes the MLH Fellowship stand out from simply contributing to large open-source projects on your own. You get personal mentorship, so you’re not stuck figuring things out alone. It makes things feel much less overwhelming, especially when working on something as massive as Airflow.
They also set up a weekly one-hour meeting every Friday called “Airflow Office Hours,” where we can ask the maintainers questions and get help with any issues in a live, synchronous call. I think this is going to be super helpful going forward.
Getting Familiar with the Codebase
The three of us spent a good chunk of time getting familiar with the product, setting up our development environments, and navigating the codebase a bit. Let me tell you, the codebase is massive! Apache Airflow is used by enterprises worldwide by their data engineering teams.
The maintainers warned us that the project can feel intimidating at first. One maintainer shared how his very first pull request took two whole months to get merged because of the complex codebase and many components. To help us get comfortable, he’s planning to walk us through how he handles an issue from start to finish so we can learn the full workflow.
Instead of diving headfirst into the code, the better approach is to first be a user and really understand how the product works. So our task for the next week is to use Airflow as much as possible, watch tutorials, read through the official docs, and try to visualize how things fit together by looking at some issues. We’ll also use tools like GitHub Copilot and Cursor to help us get a feel for the code flow.
Wrapping up
That was a short blog since I did not have much to cover but I hope you enjoyed it and stay tuned for the next blog in this series.