Introducing enhanced AI help in Amazon SageMaker Unified Studio: Agentic chat, Amazon Q Developer CLI, and MCP integration

Amazon Q Developer offers generative AI help inside Amazon SageMaker Unified Studio for knowledge discovery, knowledge processing, SQL analytics, and machine studying workflows. Right this moment, we’re asserting enhancements to the Amazon Q Developer chat expertise in SageMaker Unified Studio JupyterLab built-in improvement setting (IDE) and including Amazon Q Developer within the command line in JupyterLab and Code Editor IDEs. By integrating with Mannequin Context Protocol (MCP) servers, Amazon Q Developer is conscious of your SageMaker Unified Studio challenge sources, together with knowledge, compute, and code, and offers customized, related responses for knowledge engineering and machine studying improvement. You should utilize this improved AI help to setup your improvement setting extra shortly, and for duties like code refactoring, file modification, and troubleshooting whereas sustaining transparency into how the AI assistant is performing in your behalf.

Answer implementation

On this submit, we’ll stroll by way of how you need to use the improved Amazon Q Developer chat and the brand new built-in Amazon Q Developer CLI in SageMaker Unified Studio for coding ETL duties, to repair code errors, and generate ML improvement workflows. Each interfaces use MCP to learn recordsdata, run instructions, and work together with AWS companies instantly from the IDE. You may as well configure extra MCP servers to increase Amazon Q Developer’s capabilities with customized instruments and integrations particular to your workflow.

Stipulations

Earlier than beginning this tutorial, you will need to have the next stipulations:

• Entry to a SageMaker Unified Studio area. If you happen to don’t have a Unified Studio area, you possibly can create one utilizing the fast setup or handbook setup choice.
• Entry to or can create a SageMaker Unified Studio challenge with the All capabilities challenge profile enabled.
• Entry to or can create a JupyterLab or Code Editor compute area. We are going to stroll by way of a JupyterLab IDE instance. There isn’t any minimal occasion kind requirement to make use of the brand new options. On this submit, we use an ml.t3.medium occasion. At launch, SageMaker Distribution photographs 2.9 (comprises Amazon Q Developer chat and Amazon Q Developer CLI) or 3.4 (comprises Amazon Q Developer CLI) are required.

Importing the dataset to an Amazon S3 bucket

1. Obtain the Diabetes 130-US hospitals dataset. This dataset comprises 10 years (1999–2008) of medical care knowledge from 130 US hospitals and built-in supply networks.
2. On the Knowledge part in the midst of your challenge web page, select + on the highest. This opens Add knowledge on the appropriate.
3. On Add knowledge, select Create desk.
4. Choose Select file or drag and drop the diabetic_data CSV file.
5. Choose S3/exterior desk and full the data within the kind.
6. Choose Subsequent to add the dataset.

Amazon Q Developer chat

Amazon Q Developer chat in SageMaker Unified Studio is an agentic AI assistant that routinely understands your challenge, together with knowledge, compute sources, and code to supply extremely related strategies and insights. It helps you reply questions on your challenge, perceive advanced datasets, write code, and create notebooks, making it a strong coding companion for creating ETL workflows, constructing ML fashions, or growing generative AI functions. We are going to stroll by way of consumer personas, knowledge engineer and ML engineer, to indicate use the Amazon Q Developer chat to do exploratory knowledge evaluation, troubleshoot code, and carry out predictive evaluation. Observe: Amazon Q Developer code safety scanning will auto-scan the code as it’s being written within the IDE and supply suggestions for remediation and in some instances a code repair as properly. This helps you proactively establish and take away safety vulnerabilities in your codebase, each in present codebase and in new code as you write it within the IDE.

To launch Amazon Q Developer chat:

1. Navigate to your challenge. Entry the JupyterLab IDE. On the time of launch, Amazon Q Developer chat is simply accessible within the JupyterLab IDE.
2. Select the icon on the left for Amazon Q Developer chat. If that is the primary time opening, a message shows so that you can acknowledge the AWS insurance policies for accountable AI.
3. Enter the inquiries to work together with Amazon Q Developer chat. Enter over the Ask a query… line.

Configure extra MCP servers

You’ll be able to add extra MCP servers such because the Amazon Datazone MCP server or the AWS Knowledge Processing MCP Server to be used in Amazon Q Developer chat and the Amazon Q Developer CLI. Within the following steps, we add the AWS Knowledge Processing MCP Server, an open supply device that makes use of MCP to simplify analytics setting setup. The AWS Knowledge Processing MCP Server consists of entry to AWS Glue job statuses, Amazon Athena question outcomes, Amazon EMR cluster metrics, and AWS Glue Knowledge Catalog metadata. For extra info on configuring MCP servers, see MCP configuration for Q Developer within the IDE.

The next are the steps to configure extra MCP servers:

1. Navigate to Amazon Q Developer chat and choose the Configure MCP servers instruments icon within the higher proper. You even have the choice edit the configuration file situated at /residence/sagemaker-user/.aws/amazonq/brokers/default.json so as to add an MCP sever in Amazon Q Developer chat. You may as well navigate to /residence/sagemaker-user/.aws/amazonq/mcp.json within the terminal and edit the configuration file so as to add an MCP server in Amazon Q Developer CLI.
   
2. Choose the + image to Add new MCP server.
3. Add the next info within the kind:
4. Choose the scope: World
5. Title: Enter awsdp-mcp
6. Transport: Choose stdio
7. Command: Enteruvx
8. Arguments-optional: Enter awslabs.aws-dataprocessing-mcp-server@newest
   
9. Select Save.

Knowledge engineer

As an information engineer, you would possibly construct ETL jobs and knowledge pipelines. Amazon Q Developer chat helps cut back setup time and improves workflow effectivity by refactoring code, implementing greatest practices, and troubleshooting errors. Amazon Q Developer makes use of AI to supply code suggestions, and that is non-deterministic. The outcomes you get may be completely different from those proven within the following examples. Instance immediate:

You're a knowledge engineer. Your duty is to carry out descriptive and exploratory knowledge evaluation.
* Use the diabetic_data dataset in SageMaker Lakehouse.
* Discover checklist of connections and word down their names
* Create a pocket book. Use getting_started.ipynb for greatest practices and for instance pocket book.
* Be certain that to make use of appropriate connection names in cell magic instructions
* Be certain that to deal with lacking values, carry out descriptive evaluation, and have evaluation.
* Create a complete README.md file.
* Create a brand new working listing below the /src listing.

Run the next steps, after the answer is created.

1. Go to the pocket book.
2. Run the created pocket book and assessment every part:
   • Knowledge loading
   • Descriptive evaluation
   • Correlation matrix
   • Knowledge preprocessing similar to dealing with lacking values
   • Analyze significance of options
3. Assessment the README.md file.
4. You can also make modifications on the created recordsdata.
5. You’ll be able to immediate the Amazon Q Developer chat to make extra modifications for you.

Repair errors with out specifying the error

You can provide directions in a conversational option to Amazon Q Developer chat. With out the necessity to specify the error, Amazon Q Developer chat will entry your pocket book and repair the error.

1. Open your pocket book.
2. Immediate The pocket book isn’t working, are you able to repair it? Amazon Q Developer chat will establish the error from the pocket book.
3. Assessment the problem and the answer. Run the pocket book once more.

ML engineer

As an ML engineer, you would possibly analyze advanced datasets and run ML experiments. You’ll be able to ask Amazon Q Developer chat to tackle an ML engineer position and carry out a predictive ML mannequin on the dataset. Additionally, you possibly can ask to take the output from the information engineer into consideration. Instance immediate:

You're a machine studying engineer. Your duty is to carry out predictive machine studying mannequin on the information. The information engineer carried out exploratory evaluation. Use the output from the information engineer in your pocket book. 
- Create a pocket book to construct a diabetes prediction mannequin utilizing Amazon SageMaker.
- Be certain that to have mannequin analysis.
- Clarify your alternative for options and mannequin choice.
- Create a complete README.md file
- Do that within the working listing you created

Run the next steps, after the answer is created:

1. Run the created pocket book and assessment every part:
   • Observe that the pocket book is working efficiently.
   • Amazon Q chat included function engineering part primarily based on knowledge engineer’s output.
2. 4 ML fashions (Logistic Regression, Random Forest, Gradient Boosting, and XGBoost) had been recognized for diabetes readmission prediction.
3. Fashions had been evaluated utilizing a complete metrics suite together with accuracy, precision, recall, F1 rating, and ROC AUC to assist guarantee balanced efficiency.
4. Function engineering produced vital predictors similar to earlier inpatient visits and medicine modifications, whereas hyperparameter tuning optimized mannequin efficiency.
5. The ultimate implementation balances predictive energy with medical interpretability, enabling efficient identification of high-risk sufferers.

Amazon Q Developer CLI

The Amazon Q Developer CLI additionally understands your code, knowledge, and compute sources, however is optimized for customers preferring working within the terminal. It helps you execute and automate knowledge processing, mannequin coaching, and generative AI duties by way of pure language prompts.To launch the Amazon Q Developer CLI:

1. On the highest menu of your SageMaker Unified Studio challenge web page, select Construct, and below IDE & APPLICATIONS, select JupyterLab.
2. Look forward to the area to be prepared.
3. From the Launcher tab, open a brand new terminal. Or navigate to File > New > Terminal.
4. Enter q chat

At launch, Anthropic’s Claude Sonnet 4 in Amazon Bedrock is the default giant language mannequin (LLM). You’ll be able to select different LLMs, relying in your AWS Area. To view the accessible fashions or change the fashions enter /mannequin. MCP instruments are executable features that MCP servers expose to the Amazon Q Developer CLI. They permit Amazon Q Developer to carry out actions, course of knowledge, and work together with exterior methods in your behalf. To view the accessible instruments, enter /instruments.

Instance immediate:

Discover the datasets accessible within the challenge’s knowledge catalog and do exploratory evaluation.

Clear up

SageMaker Unified Studio by default shuts down idle sources similar to JupyterLab and Code Editor areas after 1 hour. Nonetheless, you have to delete the Amazon Easy Storage Service (Amazon S3) bucket to cease incurring extra fees. You’ll be able to delete any real-time endpoints you created utilizing the SageMaker console. For directions, see Delete Endpoints and Sources.

Conclusion

The improved AI help accessible in JupyterLab and Code Editor IDEs in SageMaker Unified Studio helps streamline knowledge engineering and machine studying workflows by offering solutions related to your challenge recordsdata, notebooks, knowledge, and compute. Whether or not you’re an information engineer constructing ETL pipelines, an information scientist conducting exploratory evaluation, or an ML engineer growing predictive fashions, these options now perceive what you’re engaged on and enable you do it extra effectively. That is simply the beginning of our agentic journey in SageMaker Unified Studio. To study extra, assessment the SageMaker Unified Studio Consumer Information. We encourage you to discover the MCP capabilities and the AWS MCP Servers repository on GitHub.

Concerning the authors

Lauren Mullennex is a Senior GenAI/ML Specialist Options Architect at AWS. She has over a decade of expertise in ML, DevOps, and infrastructure. She is a printed creator of a ebook on laptop imaginative and prescient. Exterior of labor, yow will discover her touring and mountaineering along with her two canine.

Siddharth Gupta is heading Generative AI inside SageMaker’s Unified Experiences. His focus is on driving agentic experiences, the place AI methods act autonomously on behalf of customers to perform advanced duties. Beforehand, he led edge machine studying options at AWS. This cutting-edge work goals to revolutionize how builders and knowledge scientists work together with AI, creating extra intuitive knowledge integrations and highly effective instruments for constructing and deploying machine studying fashions. An alumnus of the College of Illinois at Urbana-Champaign, he brings in depth expertise from his roles at Yahoo, Glassdoor, and Twitch. You’ll be able to attain out to him on LinkedIn.

Ishneet Kaur is a Software program Improvement Supervisor on the Amazon SageMaker Unified Studio group. She leads the engineering group to design and construct GenAI capabilities in SageMaker Unified Studio

Mohan Gandhi is a Senior Software program Engineer at AWS. He has been with AWS for the final 10 years and has labored on varied AWS companies like Amazon EMR, Amazon EFA, and Amazon RDS. Presently, he’s targeted on bettering the SageMaker inference expertise. In his spare time, he enjoys mountaineering and marathons.

Mukul Prasad is a Senior Utilized Science Supervisor within the AWS Agentic AI group. He leads the Knowledge Processing Brokers Science group growing DevOps brokers to simplify and optimize the client journey in utilizing AWS Large Knowledge processing companies together with Amazon EMR, AWS Glue, and Amazon SageMaker Unified Studio. Exterior of labor, Mukul enjoys meals, journey, images, and Cricket.

Murali Narayanaswamy is a Principal Machine Studying Scientist within the Agentic AI group in AWS engaged on merchandise together with Amazon Bedrock, Amazon SageMaker Unified Studio, Amazon Redshift and Amazon RDS. His analysis pursuits lie on the intersection of AI, optimization, studying and inference notably utilizing them to grasp, mannequin and fight noise and uncertainty in actual world functions and Reinforcement Studying in observe and at scale. Broadly, he works on utilizing concepts from on-line algorithms, optimization below uncertainty, management concept, recreation concept, synthetic intelligence, graphical fashions and estimation concept to resolve necessary issues at Amazon scale.

Necibe Ahat is a Senior AI/ML Specialist Options Architect at AWS, working with Healthcare and Life Sciences clients. Necibe helps clients to advance their generative AI and machine studying journey. She has a background in laptop science with 15 years of business expertise serving to clients ideate, design, construct and deploy options at scale. She is a passionate inclusion and variety advocate.

Vipin Mohan is a Principal Product Supervisor at Amazon Internet Companies, the place he leads generative AI product technique. He makes a speciality of constructing AI/ML merchandise, container platforms, and search applied sciences that serve 1000’s of shoppers. Exterior of labor, he mentors aspiring product managers, enjoys studying about monetary investing and entrepreneurship, and loves exploring the world by way of the eyes of his two children.