During my internship at Dagster Labs, most of my day-to-day work happened in the open — Dagster is an open source data orchestrator, and the main repo is public. That meant my code review comments, commit history, and mistakes were all visible to the world. One contribution I'm proud of is dagster-io/dagster#26365.
Open source review is a different game
In a private repo, a sloppy PR description costs you a Slack message. In a public repo, your PR description is documentation forever. Maintainers pushed me to explain why a change exists, link the issue it fixes, and include before/after behavior — not just what the diff does. That habit followed me back into closed-source work.
Tests are the conversation
The fastest way to get an open source PR merged is to make the reviewer's job trivial. A test that fails before your change and passes after it answers most review questions before they're asked. When I skipped this, reviews stalled; when I led with the failing test, they moved.
Small PRs win
A 100-line PR gets reviewed in a day. A 1000-line PR gets reviewed never.
I learned to split work aggressively: one PR for the refactor that makes the change possible, one PR for the change itself. Each is boring on its own — which is exactly what makes them easy to merge.
If you're a student
Contributing to a real open source project is the closest thing to a portable engineering credential. Interviewers can read your actual code, your actual communication, and how you handle feedback. Start with a "good first issue" on a tool you already use, and lead with a failing test.