Contributing to open source genuinely changed the trajectory of my career.

Not because of some grand contribution — because of a typo fix.

My first open source PR was fixing a typo in the README of a popular ML library. Tiny change. But it got me comfortable with the contribution process, and I started doing more: adding tests, fixing small bugs, then eventually contributing features.

With 176+ repos and 8 PyPI packages, here's what I know for sure: open source contributions on your GitHub profile carry more weight than almost anything else in ML hiring.

It shows you can navigate large codebases (most ML engineers can't). It shows you write code that meets real quality standards. It shows you collaborate with other engineers. And it's publicly verifiable — no way to fake it.

Where to start: Hugging Face Transformers has great "good first issue" labels. LangChain is rapidly growing and welcomes contributors. scikit-learn is beginner-friendly. FastAPI has a clean codebase.

Start small. Fix a typo. Add a test case. Update documentation. Then work your way up to features.

The first PR is the hardest. After that, it becomes addictive.