All news with #amazon sagemaker ai tag

🔗 Amazon SageMaker Unified Studio now supports ODBC connections, enabling Microsoft Power BI and other ODBC-compatible tools to connect directly to governed data using the Amazon Athena ODBC driver. The project overview page adds an ODBC connection details view alongside the existing JDBC view, providing DSN-less connection strings or parameters for named DSNs. The Athena ODBC driver (v2.2.0.1+) supports SageMakerBrowserIdc and SageMakerIam authentication modes. ODBC access is available today in all Regions where SageMaker Unified Studio is offered, at no additional cost.

🛠️ Amazon Timestream for InfluxDB now supports running custom Python plugins on managed InfluxDB 3 Core and Enterprise editions. You store plugin code in public or private repositories you control, and the processing engine fetches and executes it in response to supported triggers, enabling in-database data transformation, alerting, aggregation, and service integrations. Plugins run in a managed Python environment with the standard library and Amazon-vetted packages; private repos are authenticated via tokens in AWS Secrets Manager. To enable a plugin, configure a plugin repository in a DB parameter group, apply it to your cluster, and create triggers referencing the plugin via the influxdb3 CLI or HTTP API; the feature is available in all Regions where the service is offered.

🔧 Amazon SageMaker Feature Store now supports feature-level writes, enabling updates to individual features within a record without rewriting the entire record. This reduces write latency and cost by allowing pipelines to replace read-modify-write patterns with single update calls. Multiple pipelines can independently update different features in the same feature group, simplifying ingestion logic for streaming and batch jobs. The feature is available in all Regions where SageMaker Feature Store is offered.

🚀 Amazon SageMaker AI now supports Amazon EC2 G6e instances for Batch Transform, enabling high-performance offline inference on large datasets stored in Amazon S3. G6e instances feature up to eight NVIDIA L40S Tensor Core GPUs with 48 GB memory per GPU and third-generation AMD EPYC processors, improving throughput for GPU-intensive workloads. Use supported ml.g6e instance types when creating Batch Transform jobs via AWS SDKs, AWS CLI, or the CreateTransformJob API. G6e support is available in select AWS regions including US East (N. Virginia) and US West (Oregon).

🐍 Amazon SageMaker Unified Studio Workflows now includes PythonOperator and BashOperator, allowing you to run custom Python functions and shell commands inside serverless workflows without provisioning separate compute. Configure operators by adding your script files in workflow settings and pointing each operator to the function or command. This feature is available in all Regions where SageMaker Unified Studio runs.

🤖 Amazon SageMaker Unified Studio CI/CD adds two open-source features: an AI agent skill to auto-generate deployment manifests and native notebook promotion across environments. The manifest skill inspects project resources, applies least-privilege IAM guidance, and substitutes environment variables to avoid hardcoded identifiers. Notebook promotion synchronizes notebooks in place, preserving run history, supports selective promotion and dry-run validation, and integrates with existing CI/CD commands. Both features are available in all Regions offering SageMaker Unified Studio.

🧭 TabFM is a new pre-trained foundation model in BigQuery for regression and classification on tabular data, now available in preview. Exposed through SQL functions AI.PREDICT and AI.EVALUATE, it delivers instant predictions and evaluation without explicit training or feature engineering. TabFM uses in-context learning and distributed inference to handle large-scale prediction workloads and outperforms many classic models on benchmarks.

🔐 SageMaker MLflow now supports customer-managed keys (CMK) via AWS Key Management Service (KMS). This enhancement lets organizations with strict security or compliance needs manage encryption keys themselves and gain enhanced control and auditing through AWS CloudTrail. Customer-managed keys must be symmetric and created in the same AWS account and region as the MLflow App. The feature is generally available in all Regions where MLflow App is offered.

🛠️ Amazon SageMaker HyperPod now integrates Ray with built-in observability, resilient distributed training, accelerated inference, and managed development environments. Data scientists can create and manage Ray clusters from SageMaker Studio, attach JupyterLab or a local IDE for interactive iteration, and use Grafana and Amazon Managed Service for Prometheus for one-click observability. HyperPod provides node auto-recovery, hung job detection, tiered checkpointing, and task governance to improve GPU utilization and reliability, plus a tiered KV cache and JumpStart model deployment for faster Ray Serve inference.

🚀 SageMaker AI Studio now offers Generative AI Inference Recommendations, providing a guided low-code/no-code workflow to identify optimal inference configurations for generative workloads. The feature builds on an April 2026 API launch and benchmarks candidate setups on real GPU infrastructure using NVIDIA AIPerf, applying techniques like speculative decoding and kernel tuning. Users pick a use-case profile, optimization goal, and model source, then receive ranked, production-ready recommendations that can be deployed directly to SageMaker endpoints, with only standard compute costs for benchmarking.

🧭 Amazon SageMaker Notebooks now support Trusted Identity Propagation (TIP) with Amazon Athena, Amazon Redshift, and Amazon EMR Serverless, enabling per-user access control for data analytics. When connected to a TIP-enabled compute in a TIP-enabled Project, each notebook user's IAM Identity Center identity flows through to AWS Lake Formation, ensuring they see only the tables, columns, and rows their permissions allow. TIP provides per-user data boundaries, full audit attribution with CloudTrail, and reduces admin friction by automatically propagating identity through existing compute connections without extra logins or role management. The feature is available in all Regions where Amazon SageMaker Unified Studio is available.

🔍 Amazon SageMaker Unified Studio now integrates data profiling and anomaly detection powered by AWS Glue Data Quality. Data stewards, engineers, and analysts can generate dataset- and column-level statistics on catalog tables and Visual ETL job results to understand data shape and completeness. A dedicated Data profile tab supports on-demand and scheduled profiling while anomaly detection flags drift without predefined thresholds. These capabilities are available in all Regions where SageMaker Unified Studio is offered.

🔔 Amazon announced general availability of Amazon Nova Multimodal Embeddings in AWS GovCloud (US-West). The unified embedding model supports text, documents, images, video, and audio through a single model to enable cross-modal retrieval and agentic RAG. It accepts up to 8K tokens and video/audio segments up to 30 seconds, with synchronous and asynchronous API options for latency-sensitive and bulk workloads.

🔍 LocateAnything-3B, Qwen-AgentWorld-35B-A3B, and Qwen3.5-122B-A10B are now available on Amazon SageMaker JumpStart. These models provide visual grounding, agent environment simulation, and large-scale multimodal reasoning capabilities. Customers can deploy them with a few clicks via the SageMaker console or programmatically using the SageMaker Python SDK. The models expand foundation model choices for enterprise AI on AWS.

🚀 NVIDIA Nemotron 3.5 Lightning is now available on Amazon SageMaker JumpStart, enabling customers to deploy a high-throughput open model optimized for persistent agent workloads. The 30B-parameter hybrid MoE design activates 3B parameters per pass, delivering up to 4x throughput (~410 tokens/sec) and 30% faster task completion. It supports up to 1M-token context and can be post-trained and deployed across edge, on-premises, or cloud environments.

🧭 AWS Glue now lets users open Amazon SageMaker Unified Studio with a single click from the Glue console, enabling analysts and engineers to query cataloged data, run data quality checks, and build pipelines without switching contexts. Access from S3 Tables, Athena, EMR, Redshift and Glue consoles is supported, and the same IAM role is used when launching SageMaker Notebooks. An inline permissions panel simplifies setup by letting you create and customize required IAM policies in-context if SageMaker Unified Studio isn't yet configured.

🔔 Amazon SageMaker JumpStart now includes Black Forest Labs' FLUX.2-small-decoder and Google's gemma-4-12B-it, expanding foundation model options for AWS customers. FLUX.2-small-decoder offers faster image decoding with lower VRAM use, while gemma-4-12B-it provides unified multimodal understanding across text, image, and audio. Customers can deploy these models via the SageMaker console or the SageMaker Python SDK for scalable AI solutions.

🔍 Redis's langcache-embed-v3-small, JetBrains' Mellum2-12B-A2.5B-Thinking, and LightOn's LightOnOCR-2-1B are now available on Amazon SageMaker JumpStart. These models support semantic caching, code-focused reasoning, and end-to-end document OCR respectively, enabling scalable deployment on AWS. Customers can deploy them via the SageMaker JumpStart catalog or the SageMaker Python SDK with minimal effort.

🆕 Amazon SageMaker JumpStart now offers three foundation models: Z.ai’s GLM-5.2 FP8, NVIDIA’s Nemotron-Nano-12B-v2, and Z.ai’s GLM-OCR. These models cover long-horizon agentic engineering, efficient hybrid reasoning, and advanced document understanding, enabling customers to deploy high-performance AI on AWS with minimal setup. Each model is optimized for specific enterprise workflows and can be deployed via the SageMaker console or SDK.

🚀 Amazon EMR on EC2 now supports interactive Apache Spark sessions via Spark Connect, enabling data engineers and scientists to develop and debug Spark applications from managed notebooks in Amazon SageMaker Unified Studio or from IDEs like Jupyter and VS Code. Each interactive session runs on a dedicated EMR on EC2 cluster, and users can monitor and debug active and completed sessions through the EMR console. The feature offers persistent Spark contexts across cells, a client-server architecture that decouples clients from the Spark driver, and real-time observability through the Spark UI and History Server.