This week we welcome Adrin Jalali (@adrinjalali) as our PyDev of the Week! Adrin works on the popular scikit-learn package as well as Fairlearn, an AI package for Python. You can see what else Adrin is up to via his website or by checking out his Github profile.
Let’s spend some time getting to know Adrin better!
Can you tell us a little about yourself (hobbies, education, etc):
I did computer science and then bioinformatics for my master’s and PhD (which is yet to be defended!). I’m born and raised in Iran, then moved to Canada, and now live in Germany. Things I read include (leftist) economics, philosophy, diversity and inclusion, leadership, and ethics/fairness. I love cycling outside the city, and staying up all night with friends on a weekend.
Why did you start using Python?
I used to use R as a bioinformatician, and was maintaining two packages. Then I somehow studied Python over a weekend, and started coding in Python. It felt so natural that I stayed with the language. That was 2012.
What other programming languages do you know and which is your favorite?
I started with QBASIC, then Pascal, Delphi, C, C++, C#, Matlab/Octave, and R.
C++ was by far my most favourite language for a long while. I guess now it’s Python and C. I’m curious about Rust though.
What projects are you working on now?
I’ve been working mostly on scikit-learn, a machine learning library, and a bit on fairlearn which is a fariness in ML library. I also help out with the Opt Out Tool project, which is helping female identifying people to hide misogyny from their twitter feed.
Which Python libraries are your favorite (core or 3rd party)?
I’d say functools, numpy, and scipy.
How did you get involved with the scikit-learn project?
Two or three years ago we gave a talk at the PyData Berlin conference on “my first open-source contribution”. Back then I had contributed to a few different projects whenever I encountered an issue with the documentation or the code. That talk motivated me to contribute more. I tried tensorflow, but soon realized that’s more like a commercial product and not a community-based project. Then moved to scikit-learn, and immediately fell in love with the community. I have to say the mentorship I received on my GitHub pull requests especially by Joel Nothman was phenomenal and made me stay there. A few months later I was offered to be a core developer and now I’m still there.
What is your favorite feature of scikit-learn?
I love the modularity of the library. I can’t do w/o the Pipeline, *GridSearchCV, ColumnTransformer, and the fact that writing a
custom estimator is really easy. Once you’re comfortable with these mechanisms, then you can build anything you need with it.
Thanks for doing the interview, Adrin!