Machine learning in real life is hard. Finding your way into all the methods and tools is not easy. Here you will find some resources that me recently to do data versioning, tests or data drift detection.
A great comparison between all the tools you can use to do data versioning.
Managing data versions is a necessary step for data science teams to avoid output inconsistencies. Without data versioning tools, your on-call data scientist might find themselves up at 3 a.m. debugging a model issue resulting from inconsistent model outputs.
A short but useful article about tips to do tests with Pytest for Python:
- Show stdout output using pytest –capture=tee-sys
- Run single test using pytest path/to/test.py -k ‘test_foo’
Failures in data science are silent. This paper tests different solutions to get a practical way to fail loudly by doing data drift detection. The library Alibi-detect is based on this one.
This week, I published an article about The Art of being right written by Arthur Schopenhauer.
Have you ever struggled to explain why automated tests are important (or anything like that) to very reluctant and dogmatic but influential people such as well-established but very old school architects for instance? This book can help you with that.
Si vous êtes freelance ou que vous projetez de le devenir, cet épisode est pour vous. Même si vous n’êtes pas freelance, vous apprendrez quelques astuces pour négocier votre futur salaire.
If you’re a freelancer or plan to be one, this episode is for you. Even if you’re not a freelancer, you will learn some tips to negotiate your future salary.
Thank you for reading. Feel free to contact me on Twitter if you want to discuss that.