|
In the present day, we’re asserting the subsequent technology of Amazon SageMaker, a unified platform for information, analytics, and AI. The all-new SageMaker contains just about the entire parts you want for information exploration, preparation and integration, massive information processing, quick SQL analytics, machine studying (ML) mannequin growth and coaching, and generative AI software growth.
The present Amazon SageMaker has been renamed to Amazon SageMaker AI. SageMaker AI is built-in inside the subsequent technology of SageMaker whereas additionally being accessible as a standalone service for individuals who want to focus particularly on constructing, coaching, and deploying AI and ML fashions at scale.
Highlights of the brand new Amazon SageMaker
At its core is SageMaker Unified Studio (preview), a single information and AI growth surroundings. It brings collectively performance and instruments from the vary of standalone “studios,” question editors, and visible instruments that we now have right now in Amazon Athena, Amazon EMR, AWS Glue, Amazon Redshift, Amazon Managed Workflows for Apache Airflow (MWAA), and the prevailing SageMaker Studio. We’ve additionally built-in Amazon Bedrock IDE (preview), an up to date model of Amazon Bedrock Studio, to construct and customise generative AI purposes. As well as, Amazon Q supplies AI help all through your workflows in SageMaker.
Right here’s an inventory of key capabilities:
- Amazon SageMaker Unified Studio (preview) – Construct with all of your information and instruments for analytics and AI in a single surroundings.
- Amazon SageMaker Lakehouse – Unify information throughout Amazon Easy Storage Service (Amazon S3) information lakes, Amazon Redshift information warehouses, and third-party and federated information sources with Amazon SageMaker Lakehouse.
- Information and AI Governance – Securely uncover, govern, and collaborate on information and AI with Amazon SageMaker Catalog, constructed on Amazon DataZone.
- Information Processing – Analyze, put together, and combine information for analytics and AI utilizing open supply frameworks on Amazon Athena, Amazon EMR, and AWS Glue.
- Mannequin growth – Construct, prepare, and deploy ML and basis fashions (FMs) with absolutely managed infrastructure, instruments, and workflows with Amazon SageMaker AI.
- Generative AI app growth – Construct and scale generative AI purposes with Amazon Bedrock.
- SQL analytics – Achieve insights with Amazon Redshift, probably the most price-performant SQL engine.
On this put up, I provide you with a fast tour of the brand new SageMaker Unified Studio expertise and how one can get began with information processing, mannequin growth, and generative AI app growth.
Working with Amazon SageMaker Unified Studio (preview)
With SageMaker Unified Studio, you possibly can uncover your information and put it to work utilizing acquainted AWS instruments to finish end-to-end growth workflows, together with information evaluation, information processing, mannequin coaching, and generative AI app constructing, in a single ruled surroundings.
An built-in SQL editor helps you to question information from a number of sources, and a visible extract, rework, and cargo (ETL) device simplifies the creation of information integration and transformation workflows. New unified Jupyter notebooks allow seamless work throughout totally different compute companies and clusters. With the brand new built-in information catalog performance, yow will discover, entry, and question information and AI belongings throughout your group. Amazon Q is built-in to streamline duties throughout the event lifecycle.
Let’s discover the person capabilities in additional element.
Information processing
SageMaker integrates with SageMaker Lakehouse and allows you to analyze, put together, combine, and orchestrate your information in a unified expertise. You may combine and course of information from varied sources utilizing the supplied connectivity choices.
Begin by making a challenge in SageMaker Unified Studio, selecting the SQL analytics or information analytics and AI-ML mannequin growth challenge profile. Tasks are a spot to collaborate along with your colleagues, share information, and use instruments to work with information in a safe method. Undertaking profiles in SageMaker outline the preconfigured set of sources and instruments which are provisioned while you create a brand new challenge. In your challenge, select Information within the left menu and begin including information sources.
The built-in SQL question editor helps you to question your information saved in information lakes, information warehouses, databases, and purposes straight inside SageMaker Unified Studio. Within the prime menu of SageMaker Unified Studio, choose Construct and select Question Editor to get began. Additionally, strive creating SQL queries utilizing pure language with Amazon Q whilst you’re at it.
You also needs to discover the built-in visible ETL device to create information integration and transformation workflows utilizing a visible, drag-and-drop interface. Within the prime menu, choose Construct and select Visible ETL circulate to get began.
If Amazon Q is enabled, you can too use generative AI to creator flows. Visible ETL comes with a variety of information connectors, pre-built transformations, and options reminiscent of scheduling, monitoring, and information previewing to streamline your information workflows.
Mannequin growth
SageMaker Unified Studio contains capabilities from SageMaker AI, which supplies infrastructure, instruments, and workflows for the whole ML lifecycle. From the highest menu, choose Construct to entry instruments for information preparation, mannequin coaching, experiment monitoring, pipeline creation, and orchestration. It’s also possible to use these instruments for mannequin deployment and inference, machine studying operations (MLOps) implementation, mannequin monitoring and analysis, in addition to governance and compliance.
To begin your mannequin growth, create a challenge in SageMaker Unified Studio utilizing the information analytics and AI-ML mannequin growth challenge profile and discover the brand new unified Jupyter notebooks. Within the prime menu, choose Construct and select JupyterLab. You need to use the brand new unified notebooks to seamlessly work throughout totally different compute companies and clusters. You need to use these notebooks to modify between environments with out leaving your workspace, streamlining your mannequin growth course of.
It’s also possible to use Amazon Q Developer to help with duties reminiscent of code technology, debugging, and optimization all through your mannequin growth course of.
Generative AI app growth
Use the brand new Amazon Bedrock IDE to develop generative AI purposes inside Amazon SageMaker Unified Studio. The Amazon Bedrock IDE contains instruments to construct and customise generative AI purposes utilizing FMs and superior capabilities reminiscent of Amazon Bedrock Data Bases, Amazon Bedrock Guardrails, Amazon Bedrock Brokers, and Amazon Bedrock Flows to create tailor-made options aligned along with your necessities and accountable AI pointers.
Select Uncover within the prime menu of SageMaker Unified Studio to browse Amazon Bedrock fashions or experiment with the mannequin playgrounds.
Create a challenge utilizing the GenAI Utility Growth profile to start out constructing generative AI purposes. Select Construct within the prime menu of SageMaker Unified Studio and choose Chat agent.
With the Amazon Bedrock IDE, you possibly can construct chat brokers and create data bases out of your proprietary information sources with only a few clicks, enabling Retrieval-Augmented Era (RAG). You may add guardrails to advertise secure AI interactions and create capabilities to combine with any system. With built-in mannequin analysis options, you possibly can check and optimize your AI purposes’ efficiency whereas collaborating along with your crew. Design flows for deterministic genAI-powered workflows, and when prepared, share your purposes or prompts inside the area or export them for deployment anyplace—all whereas sustaining management of your challenge and area belongings.
For an in depth description of all Amazon SageMaker capabilities, examine the SageMaker Unified Studio Person Information.
Getting began
To start utilizing SageMaker Unified Studio, directors want to finish a number of setup steps. This contains establishing AWS IAM Identification Heart, configuring the mandatory digital personal cloud (VPC) and AWS Identification and Entry Administration (IAM) roles, making a SageMaker area, and enabling Amazon Q Developer Professional. As an alternative of IAM Identification Heart, you can too configure SAML by way of IAM federation for consumer administration.
After the surroundings is configured, customers check in by way of the supplied SageMaker Unified Studio area URL with single sign-on. You may create tasks to collaborate with crew members, selecting from pre-configured challenge profiles for various use instances. Every challenge connects to a Git repository for model management and contains an instance unified Jupyter pocket book to get you began.
For detailed setup directions, examine the SageMaker Unified Studio Administrator Information.
Now accessible
The subsequent technology of Amazon SageMaker is on the market right now within the US East (N. Virginia, Ohio), US West (Oregon), Asia Pacific (Tokyo), and Europe (Eire) AWS Areas. Amazon SageMaker Unified Studio and Amazon Bedrock IDE can be found right now in preview in these AWS Areas. Test the full Area checklist for future updates.
For pricing info, go to Amazon SageMaker pricing and Amazon Bedrock pricing. To be taught extra, go to Amazon SageMaker, SageMaker Unified Studio, and Amazon Bedrock IDE.
Present Amazon Bedrock Studio preview domains might be accessible till February 28, 2025, however it’s possible you’ll not create new workspaces. To expertise the superior options of Bedrock IDE, create a brand new SageMaker area following the directions within the Administrator Information.
Give the brand new Amazon SageMaker a strive within the console right now and tell us what you suppose! Ship suggestions to AWS re:Publish for Amazon SageMaker or by way of your standard AWS Help contacts.
— Antje
