My new job is building a data science infrastructure from the ground up at a startup. I’m tracking the how’s and why’s and plan to blog along the way, but wondering — does anyone have questions or topics they’d like to hear about along these lines? #datascience #py4ds #r4ds
— Caitlin Hudon👩🏼💻 (@beeonaposy) October 1, 2018
similar to @therriaultphd's point: DataSci DevOps is sorely lacking best practices. (DevOps understood as making every dev capable of reliably pushing to prod often, and DSDevOps understood as making every data scientist capable of reliably putting their models in prod often)
— Henri Palacci (@henripal) October 1, 2018
Good point! is the DS a core component of the product or is it sophisticated analytics?
— Nidhal Selmi - نضال السالمي (@imleslahdin) October 1, 2018
If the latter then the focus on infrastructure is quite interesting this "early".
If the former, it would be hard to write in a meaningful way without giving out details about your product.
How to properly conduct stakeholder management around expectations for experiment outcomes (results, time, engineering resources)
— Kevin Liu (@ckevinliu) October 2, 2018
