Completing the Machine Learning Loop

|

Pachyderm’s ability to version data and run pipelines at scale is at the foundation of bringing ML to software development. In a new blog post, Completing the Machine Learning Loop, we see why treating data as a first-class citizen in machine learning (ML) development is necessary to enable fast and reliable iteration on AI software. The two loops model presented in the blog, gives us a mental framework for how to apply MLOps and DevOps best practices to the two moving pieces of ML development: code and data. This understanding is essential for anyone working to deploy sustainable Ml-based systems into a real-world environments.