Book a call

News · December 8, 2020

AWS adds data prep, feature store and pipelines to SageMaker

// ai.news :: 2020

At its online re:Invent conference, AWS announced a batch of SageMaker features aimed at the unglamorous parts of machine learning: preparing data, reusing features, automating pipelines and checking for bias. It lowers the bar for running ML in production.

What happened

  • AWS announced new Amazon SageMaker capabilities at re:Invent, including Data Wrangler for preparing data with over 300 built-in transformations without code, and Feature Store for sharing features between training and inference, according to its December 8, 2020 release.
  • SageMaker Pipelines adds CI/CD-style automation for ML workflows, and SageMaker Clarify detects statistical bias and explains how features affect predictions, AWS said.
  • The release also described deeper profiling in SageMaker Debugger, distributed training that AWS said can train large models up to two times faster, Edge Manager for models on device fleets, and JumpStart with pre-built solutions and models.
  • Pure AI reported that the update was meant to help developers automate and scale each step of the ML workflow.

Why it matters

Most ML projects stall between a working notebook and a reliable production service. Managed pipelines, a feature store and bias reports address exactly that gap, and they make it realistic for smaller teams to run models without a dedicated platform group.

The trade-off is deeper lock-in to one cloud. We would adopt the pieces that remove the most manual work first, usually pipelines and monitoring, keep model code and data definitions portable, and use the bias and explainability reports as documentation that customers and regulators are starting to ask for.

Sources

This post is CMM’s own summary and commentary on publicly reported news. Facts are attributed to the sources linked above; please read them for full details. CMM is not affiliated with these publishers or with the companies mentioned, and trademarks belong to their owners. Prepared with AI assistance. Corrections or requests: [email protected]. See our editorial policy.

Want help with this?

Tell us what you need, and we will suggest the right team model on a short call.

Talk to us