This is true, and I talk in my "Becoming a Data Scientist" talk about how you're probably ready to start applying for jobs earlier than you think.https://t.co/BCP9u93n9g
— Data Science Renee (@BecomingDataSci) July 27, 2018
P.S. In data science, there's no way to "learn it all" anyway. We're all learning on the job all the time. https://t.co/nvQGnelOUQ
Hot take: `foo` and `bar` are bad variable names to use in examples. They confuse learners who (rightly) assume they have implicit meanings. Instead, use variable names in an understandable example scenario, like `sales` and `customers`.
— Chris Albon (@chrisalbon) July 26, 2018
The "no googling or looking things up in books" requirement for coding interviews and takehome assignments is bullshit. I'm constantly looking up tech stuff in books THAT I WROTE.
— Al Sweigart (@AlSweigart) July 24, 2018
The nonfiction section of any bookstore these days is a long series of "Catchy keyword: how this one weird idea explains everything about the world"...
— François Chollet (@fchollet) July 24, 2018
Anyone interested in #AutoML, especially as applied to #DeepLearning, should read this new post by @math_rachel from @fastdotai. Rachel consistently delivers thoughtful, no-BS perspectives on many topics in #MachineLearning and #DeepLearning. 👏https://t.co/NULD6qInsb
— Erin LeDell (@ledell) July 24, 2018
my new post: Is Google AutoML the key to democratizing AI?https://t.co/b5kxp3lmEw
— Rachel Thomas (@math_rachel) July 23, 2018
me coding by myself vs me during a coding phone screen pic.twitter.com/6s00MQZbaS
— Christina Zhu (@cszhu) July 22, 2018
Fewer rockstars. More mentors. https://t.co/OoccGK1lsD
— Eva (@evacide) July 21, 2018
When writing my master’s thesis I recall being told that easy_to_grok_variable_names were not how academic papers were written. I explained that it was so much easier to understand if I skipped the Greek. It was made clear to me that being understood was secondary to conformity. https://t.co/kYFPr6dlWs
— JD Long (@CMastication) July 21, 2018
The real world is not differentiable.
— hardmaru (@hardmaru) July 20, 2018
I feel so seen by @math_rachel. Great post: https://t.co/Sa7Rk2JWGr. Includes a link to one of my favorite ML papers, "Machine Learning: The High-Interest Credit Card of Technical Debt". pic.twitter.com/FQT8pfnrA3
— Vicki Boykis (@vboykis) July 19, 2018
3. But more often than you might think, the missing/miscoded/"outlier" cases indicate a larger, more systematic problem with your code or with your data. My advice is to do due diligence on those cases BEFORE moving on to next steps, because errors tend to compound.
— Nate Silver (@NateSilver538) July 18, 2018
