Key features

Connect to over 40 AWS and third-party data sources, import data, verify data quality, engineer model features using 300+ built-in data transformations, and save them to SageMaker Feature Store in a few clicks using SageMaker Data Wrangler. You can create or schedule Data Wrangler jobs to process data at scale, and automate data preparation steps in the ML workflow using SageMaker Pipelines.

Simplify your data workflows with a unified notebook environment for data engineering, analytics, and ML. Create, browse, and connect to Amazon EMR clusters and AWS Glue Interactive Sessions directly from Studio notebooks. Monitor and debug Spark jobs using familiar tools such as Spark UI right from the notebooks. Use the built-in data preparation capability powered by SageMaker Data Wrangler directly from Studio notebooks to visualize data, identify data quality issues, and apply recommended solutions to improve data quality and model accuracy without writing a single line of code.

Connect to data stores, spin up the resources to run your data processing job, save the output to persistent storage, and provide logs and metrics using SageMaker Processing.

Store, share, and manage ML model features for training and inference to promote feature reuse across ML applications using SageMaker Feature Store, a fully managed, purpose-built repository in SageMaker Studio. You get the same features consistently both during training and during inference, saving months of development effort.

Access fully managed Jupyter notebooks in SageMaker Studio in one click. SageMaker Studio notebooks come preconfigured with deep learning environments for TensorFlow and PyTorch (optimized by AWS) to help you quickly get started with model building. You can dial up or down the underlying compute resources without interrupting your work.

Coedit the same notebook file, run notebook code simultaneously, and review the results together to streamline collaboration. All resources are automatically tagged, making it easier to monitor cost and usage of SageMaker Studio.

Use over 15 built-in algorithms available in prebuilt container images to quickly train and run inference, or bring your own custom images to SageMaker Studio.

Automatically build, train, and tune the best ML models based on your data while maintaining full control and visibility using SageMaker Autopilot. Then directly deploy models to production with just one click. You can also automatically generate a SageMaker Studio notebook for any model SageMaker Autopilot creates and dive into the details of how it was created, refine it as desired, and recreate it from the notebook.

Quickly get started with ML using hundreds of prebuilt solutions that can be deployed with just a few clicks using SageMaker JumpStart.

Set up a distributed compute cluster, perform the training, output results to Amazon Simple Storage Service (S3), and tear down the cluster in a single click. Train models at scale using SageMaker data parallel and model parallel libraries, and accelerate training processes by up to 50% through graph- and kernel-level optimizations by using SageMaker Training Compiler. You can reduce costs by up to 90% by using managed spot instance training.

Track iterations to ML models by capturing the input parameters, configurations, and results, and storing them as experiments using SageMaker Experiments. You can browse active experiments, search and review previous experiments, and compare results across experiments.

Automatically tune your model by adjusting thousands of combinations of algorithm parameters to arrive at the most accurate predictions the model is capable of producing, saving weeks of effort.

Capture metrics and profile training jobs in real time so you can correct performance problems quickly, before the model is deployed to production, using SageMaker Debugger.

Deploy your trained model into production with a single click. Access SageMaker Model Deployment within SageMaker Studio for all your inference needs, from low latency (a few milliseconds) and high throughput (hundreds of thousands of requests per second) to long-running inference for use cases such as natural language processing and computer vision.

Deploy thousands of models on a single endpoint using SageMaker’s multi-model endpoints and multi-container endpoints, improving cost-effectiveness while providing the flexibility to use models as often as you need them.

Track model versions, their metadata, and performance using SageMaker Model Registry, making it easier to choose the right model for deployment based on your business requirements. In addition, you can automatically log approval workflows for audit and compliance.

Bring continuous integration and delivery (CI/CD) practices to ML using SageMaker Projects, such as maintaining parity between development and production environments, source and version control, A/B testing, and automation.

Maintain quality by detecting model drift and concept drift in real time using SageMaker Model Monitor within SageMaker Studio. All models trained in SageMaker automatically emit key metrics that can be collected and viewed in SageMaker Studio.

Once a notebook is selected, Amazon SageMaker Studio notebook takes a snapshot of the entire notebook, packages its dependencies in a container, builds the infrastructure, runs the notebook as an automated job on a schedule set by the practitioner, and deprovisions the infrastructure upon job completion–reducing the time it takes to move a notebook to production from weeks to hours.

Automate the entire model build workflow, including data preparation, feature engineering, model training, model tuning, and model validation using SageMaker Pipelines. You can configure SageMaker Pipelines to run automatically at regular intervals or when certain events occur, or run them manually as needed.