We hosted PuPPy, the Puget Sound Programming in Python meetup group, in the Redfin Seattle office yesterday — here’s a link to the slides.
Redfin has written an ETL Testing pipeline that allows us to regression test our data importers, and to make changes to our analytics Redshift cluster ETL via a continuous deployment pipeline.
We use a system codename dirpy to resize images in 6 – 8 ms to allow us to use any image from our data sources as soon as they are available.
We are rewriting our build in Bazel, a Google-built build system that allows us to do fast, correct builds, in anticipation of shipping a continuous deployment pipeline for all our code later this summer.
See more in the slides.