{"version":1,"type":"story","url":"https://digestai.news/story/posit-launches-positron-on-amazon-sagemaker-ai-for-data-science-workfl","json":"https://digestai.news/story/posit-launches-positron-on-amazon-sagemaker-ai-for-data-science-workfl.json","markdown":"https://digestai.news/story/posit-launches-positron-on-amazon-sagemaker-ai-for-data-science-workfl.md","slug":"posit-launches-positron-on-amazon-sagemaker-ai-for-data-science-workfl","headline":"Posit launches Positron on Amazon SageMaker AI for data science workflows","summary":"Posit’s IDE, Positron, is now offered as a custom container that runs inside Amazon SageMaker Studio Spaces. The image, built on the SageMaker Distribution base, lets a Space execute Athena queries, read the AWS Glue Data Catalog and access S3 storage using the Space execution role, so no keys need to be managed. Data scientists can launch a Space on any instance size, reserve capacity with SageMaker training plans, and invoke Posit Assistant, which calls Amazon Bedrock models inside the same AWS account.\n\nThe blog walks through a synthetic loan‑portfolio use case of 50,000 records stored in S3 and registered in Glue. Athena returns a 2.18 MiB sample in under a second. After cleaning, 48,500 rows are used to train an XGBoost classifier on an ml.t3.xlarge instance. The held‑out evaluation reports an AUC of 0.834 and a 12.3 % observed default rate in the highest‑risk decile, while the overall default rate is 2.03 %. Posit Assistant processed 6,657,942 tokens, estimated at $6.319 cost with 92.5 % cache efficiency. The trained model is packaged, deployed as a real‑time SageMaker endpoint, and accessed from a Shiny for Python app and a Quarto report, all within the same Positron project.\n\nThe run is illustrative only; the data and payloads are synthetic, and the blog notes that production use still requires validation of licensing, security, compliance and performance characteristics. Administrators must build and register the custom image in a private ECR repository, attach it to a Studio domain, and manage role permissions. Cleanup steps are provided to avoid lingering charges.","keyPoints":["Positron runs in SageMaker Studio Spaces via a custom container, giving Athena, Glue, and S3 access through the Space execution role.","Posit Assistant used Amazon Bedrock, processing 6,657,942 tokens at an estimated $6.319 cost with 92.5% cache efficiency.","An XGBoost classifier trained on 40,000 rows achieved 0.834 AUC and a 12.3% default rate in the top risk decile."],"whyItMatters":"Integrating Positron with SageMaker lets data teams work in a single, governed environment, eliminating credential handling and streamlining R, Python, model deployment and reporting, which speeds development and reduces operational overhead.","category":{"slug":"enterprise","name":"Enterprise & Industry","url":"https://digestai.news/category/enterprise"},"entities":{"companies":["Posit","Amazon","AWS","Amazon SageMaker","Amazon Athena","Amazon Bedrock"],"models":["XGBoost","Amazon Bedrock"],"people":[]},"firstPublishedAt":"2026-09-21T16:34:21Z","updatedAt":"2026-09-21T16:34:21Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"AWS Machine Learning Blog","title":"Run Positron on Amazon SageMaker AI for data science workflows","url":"https://aws.amazon.com/blogs/machine-learning/run-positron-on-amazon-sagemaker-ai-for-data-science-workflows","publishedAt":"2026-09-21T16:34:21Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[],"thread":{"title":"Amazon SageMaker Expands AI Inference Capabilities","url":"https://digestai.news/thread/amazon-announces-sagemaker-hyperpod-inference-gateway-to-cut-first-token","storyCount":2},"cite":{"text":"Digest AI, \"Posit launches Positron on Amazon SageMaker AI for data science workflows\", 21 September 2026, https://digestai.news/story/posit-launches-positron-on-amazon-sagemaker-ai-for-data-science-workfl","publisher":"Digest AI","title":"Posit launches Positron on Amazon SageMaker AI for data science workflows","datePublished":"2026-09-21T16:34:21Z","url":"https://digestai.news/story/posit-launches-positron-on-amazon-sagemaker-ai-for-data-science-workfl"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}