Spot the difference in ML costs (28 Jan 2021)
Spot instances are a great option for anyone training machine learning models; they're much cheaper than other on-demand options, albeit with some drawbacks.
Is your Machine Learning Reproducible? (19 Jan 2021)
Short answer: not really, but it can become better!
A most unusual year (26 Dec 2020)
A retrospective of some of the decisions that caused us to pivot from a focused machine learning consultancy to deciding to build an open-source MLOps tool.
ZenML will be open source (11 Nov 2020)
An overview of some of the capabilities that ZenML will unlock for our users.
MLOps: Learning from history (09 Nov 2020)
MLOps isn't just about new technologies and coding practices. Getting better at productionizing your models also likely requires some institutional and/or organisational shifts.
Why ML in production is (still) broken - [#MLOps2020] (26 Jun 2020)
The MLOps movement and associated new tooling is starting to help tackle the very real technical debt problems associated with machine learning in production.
Distributed PCA using TFX (27 Feb 2020)
We use PCA to reduce the dimension of input vectors while retaining maximal variance.