👥 Edge#153: ML Model Versioning
🎄 You’re on the free list for Sunday news digest TheSequence Scope. To celebrate you and our 120,000 readers, we offer a unique 60% discount for the Premium subscription to TheSequence. Subscribe and share with colleagues! Thank you.🎄
Til January 7th:
💡 ML Concept of the Day: ML Model Versioning Â
Happy start of 2022! The second part of our popular MLOPs series (Part 1 is here), we’d like to start with version control in ML pipelines. Versioning is one of those aspects that we tend to ignore until they become a problem. This is partly because versioning is rarely an issue when we talk about a handful of models but can become a total nightmare in a medium to large-scale ML infrastructure →read more with the subscription
🔎 ML Research You Should Know: How Uber Backtests and Versions Forecasting Models at ScaleÂ
Time-series forecasting is a key component of Uber’s machine learning architecture. Across its several properties, Uber runs thousands of time-series forecast models across diverse areas such as ride planning or budget management. Ensuring the accuracy of those forecast models is far from being an easy endeavor. The backtesting frameworks such as Omphalos that Uber has built previously have proven to be effective for some specific use cases but were unable to scale with Uber’s operation. →read along to know how did they address the limitations of previous efforts
🤖 ML Technology to Follow: Lyft’s Amundsen is an Open Source Data Discovery and Versioning Platform for Data Science WorkflowsÂ
Why should I know about this: Amundsen is one of the data versioning platforms that have been applied in large-scale ML pipelines →more Â