Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

COMPANY · ENTITY #9826

AWS Identity and Access Management (IAM)

Related event timeline, sources and context from the news index.

EVENT TIMELINE

19

MODELS · 1 SOURCE · AWS Machine Learning

Amazon SageMaker HyperPod adds Global Inference Gateway for cross-cluster and cross-Region routing

Amazon announced the Global Inference Gateway for SageMaker HyperPod, a SageMaker-managed gateway that routes inference requests across registered HyperPod clusters in one or multiple AWS Regions. The gateway authenticates requests with SigV4/IAM, forwards traffic over AWS PrivateLink, provides automatic failover and per-request cluster selection, and exposes a single endpoint instead of per-cluster endpoints.

7.0

COMPANIES · 1 SOURCE · AWS Machine Learning

Standard Bank automates generative AI risk assessments using Amazon Bedrock

Standard Bank Group, with AWS Professional Services, deployed an automated generative-AI assessment pipeline running in the bank's AWS account that collects evidence (docs, code, live calls), uses a foundation model on Amazon Bedrock as an "LLM-as-a-judge," and scores use cases against international standards and the bank's policies, delivering reports into their existing approval gates. The pipeline uses AWS Well-Architected Tool, Step Functions, S3, a custom 40-question lens, and a pre-assessment agent on AWS Fargate to pre-populate evidence and speed risk committee review.

6.0

COMPANIES · 1 SOURCE · AWS Machine Learning

Amazon SageMaker AI adds automatic monitoring and recommendations for inference endpoints

Amazon SageMaker AI now monitors generative-AI inference endpoints (default for vLLM serving containers on GPU instances) and flags latency drift, unused GPU capacity, and missing serving optimizations. Findings are sent via AWS Health and open an agentic experience in SageMaker Studio where teams can run benchmarked inference recommendations on separate compute and apply changes manually; the monitoring inspects operational signals only and does not read requests or responses.

6.0

MODELS · 1 SOURCE · AWS Machine Learning

PolyAI’s Dialog-RSN-1 available on Amazon SageMaker JumpStart and AWS Marketplace

PolyAI’s Dialog-RSN-1, a real-time audio-first reasoning model for live conversational agents, is now listed on AWS Marketplace and can be deployed as an Amazon SageMaker JumpStart endpoint inside customers’ own AWS accounts. The model perceives caller audio directly, reasons and responds in real time, speaks the OpenAI Realtime protocol, and is English-only today; one-click deployment and a SageMaker Python SDK path are provided.

7.0

COMPANIES · 1 SOURCE · AWS Machine Learning

Amazon SageMaker HyperPod adds enhanced GPU infrastructure observability with CloudWatch default

Amazon announced enhanced observability for SageMaker HyperPod: a new add-on/extension that provides a diagnostic summary (using CloudWatch's DetectGroupOutliers API and static diagnostic rules) and two telemetry tiers — default metrics (30s collection, enabled automatically) and opt-in advanced metrics (full node exporter, NVIDIA DCGM, AWS Neuron, EFA). The feature supports both EKS- and Slurm-orchestrated training, surfaces per-job/node/accelerator issues (XID/ECC faults, throttling, network degradation), and makes Amazon CloudWatch the new default monitoring backend alongside optional Amazon Managed Service for Prometheus.

6.0

MODELS · 1 SOURCE · AWS Machine Learning

Stretto builds Tara AI voice agent with Amazon Nova 2 Sonic (Amazon Bedrock)

Stretto, working with the AWS Generative AI Innovation Center and CI&T, built Tara — a phone-based bankruptcy intake AI voice agent — using the speech-to-speech model Amazon Nova 2 Sonic via Amazon Bedrock. Tara conducts structured intake interviews, extracts financial data, applies attorney-defined lead-sorting rules, and Stretto reports a 95% reduction in per-minute voice cost compared to its previous solution while improving latency and interruption handling.

7.0

COMPANIES · 1 SOURCE · AWS Machine Learning

Natera builds SWARM agentic AI for lab ops on Amazon Bedrock AgentCore

Natera developed SWARM, an agentic AI system for lab operations, using Amazon Bedrock AgentCore to capture and deploy institutional troubleshooting knowledge at scale. The system routes Jira tickets through EventBridge/SQS, provides two runtimes (swarm-agent and swarm-hive) backed by a shared skill library and AWS services, and reports ~130× faster first response time and 96.2% automated coverage for support tickets.

7.0

MODELS · 1 SOURCE · AWS Machine Learning

Amazon Bedrock framework to optimize cost-per-task and measure agent ROI

AWS/Amazon blog post presents a framework for calculating true model cost on Bedrock (including four hidden cost drivers) and techniques to reduce cost-per-task—caching, routing, and output constraints—plus a discipline for measuring agent ROI using human baselines, holdout experiments, and a five-dimension scorecard. It stresses cost-per-task (not price-per-token) and explains how tokenizer differences, output-token premiums, multi-turn agent loops, and reasoning tokens compound real costs (including an illustrative example where a higher-priced model costs 65% less).

7.0

CODING · 1 SOURCE · AWS Machine Learning

Building a self-service LLM gateway with Amazon Bedrock and LiteLLM

A how-to post describes a pattern that uses LiteLLM (running on AWS Fargate) to expose an OpenAI-compatible API that proxies Amazon Bedrock models, enabling per-user API keys, cost controls, and observability via CloudWatch and Langfuse. The walkthrough includes AWS CDK examples, uses Claude Sonnet and Claude Haiku as examples, and outlines storage, caching, and IAM configurations for a production topology.

5.0

CODING · 1 SOURCE · AWS Machine Learning

AWS publishes Sample Bedrock Invocation Data Lake for centralized observability

AWS published a sample implementation (GitHub repo) that centralizes Amazon Bedrock invocation logs across linked accounts into a queryable data lake using CloudWatch subscription filters, Data Firehose, and Athena. The repo includes CloudFormation templates, Lambda code, and SQL queries to capture per-invocation details such as input/output token counts, model used, stop reason, account, and IAM identity.

5.0

COMPANIES · 1 SOURCE · AWS Machine Learning

AWS launches first brand campaign for enterprise AI workspace AWS Quick

AWS is launching its first brand campaign for AWS Quick, an AI workspace built specifically for enterprise teams that integrates with existing apps and enterprise security controls. The campaign will run in the US and UK across streaming TV, online video, social media, and business press; AWS says Quick enforces permissions, supports Microsoft Purview, IAM and Active Directory, holds certifications such as HIPAA, FedRAMP, SOC 2 and ISO 27001, and keeps data and chat transcripts private and within customers' environments.

6.0

COMPANIES · 1 SOURCE · AWS Machine Learning

Amazon Bedrock AgentCore launches Web Answer (general availability)

Amazon announced general availability of Web Answer on Bedrock AgentCore, a managed connector that takes a natural-language question and returns a synthesized, source-cited answer in seconds. Web Answer runs live web searches, performs an agentic retrieval-and-verification loop, synthesizes responses with URL-level citations, and is callable via an AgentCore Gateway using Model Context Protocol (MCP).

7.0

COMPANIES · 1 SOURCE · AWS Machine Learning

ChatGPT desktop app configured to use Amazon Bedrock in AWS GovCloud (US)

OpenAI’s ChatGPT desktop app can be configured to route model inference to Amazon Bedrock Mantle in AWS GovCloud (US) (us-gov-west-1), giving access to OpenAI frontier models and the Codex coding agent while using AWS IAM, region controls, and optional zero data retention. The walkthrough covers authentication options (Bedrock API key or AWS SDK credential chain, including AWS SSO), config files, network restrictions, and a controlled local deployment workflow for organizations with strict data‑residency and security requirements.

7.0

COMPANIES · 1 SOURCE · AWS Machine Learning

Octave Reliance builds secure multi-tenant AI foundation on Amazon Bedrock with Cohere Embed v4

Octave Reliance built a three-layer, multi-tenant AI foundation on Amazon Bedrock using Cohere Embed v4 to enable secure semantic search and reasoning over regulated documents. The design scopes vector stores, retrieval, and agent execution to tenant and user boundaries, embedded roughly 2 billion documents, and delivered a Complaint Triage Assistant in production while cutting per-use-case deployment time by about 50%.

7.0

MODELS · 1 SOURCE · AWS Machine Learning

Claude Opus 5.5 and Sonnet 5.5 available on Amazon Bedrock in AWS GovCloud (US) with Claude Code

Anthropic’s Claude Opus 5.5 and Claude Sonnet 5.5 (alongside Claude Sonnet 5) are available via Amazon Bedrock in AWS GovCloud (US); Sonnet 5 holds FedRAMP Class D and DoD IL4/IL5 authorizations and Opus/Sonnet 5.5 hold FedRAMP Class D certification on Bedrock (users should verify current model certification status). The post outlines using Anthropic’s agentic coding tool Claude Code with Bedrock’s bedrock-runtime and bedrock-mantle endpoints, describing security features (Zero Operator Access), compliance pathways, and integrations for CI/CD and developer tooling.

8.0

COMPANIES · 1 SOURCE · AWS Machine Learning

Manage Amazon SageMaker HyperPod Spaces directly from SageMaker Studio

AWS added UI support in SageMaker Studio to create and manage SageMaker Spaces on HyperPod EKS clusters, enabling data scientists to launch JupyterLab or Code Editor environments and start/stop Spaces without using the CLI. Administrators must install the SageMaker Spaces add-on on HyperPod clusters, attach required IAM policies, and enable per-user identity propagation for Studio domains.

6.0

CODING · 1 SOURCE · AWS Machine Learning

Anthropic guide: implement multi-environment access for Claude Platform on AWS

Anthropic published a step‑by‑step implementation guide for Claude Platform on AWS (CPonAWS) that shows how to enable inference from three environments—AWS production workloads via cross‑account SigV4, developer laptops via workspace‑scoped API keys, and external services via OIDC federation and short‑lived tokens—using a dedicated AI Services linked account and workspace isolation. The walkthrough includes CLI commands, console instructions, code snippets, and a recommended three‑account topology (payer, AI Services, workload) with examples for IAM roles, workspace ARNs, and region‑specific endpoints.

6.0

MODELS · 1 SOURCE · AWS Machine Learning

OpenAI's GPT-6.1 Sol becomes generally available on Amazon Bedrock

According to OpenAI, GPT-6.1 Sol is now generally available on Amazon Bedrock, offering stronger reasoning for agentic coding, complex document analysis, and multistep professional workflows. OpenAI claims GPT-6.1 Sol matches GPT-6 Astra on DeepSWE v1.1 at roughly one-fifth the cost per task and outperforms GPT-6 Sol; the model is accessible via Codex, ChatGPT Work and Bedrock APIs and can be run with AWS controls (IAM, CloudTrail, VPC endpoints) and hardware-isolated inference with customer data-not-used-for-training options.

8.0

MODELS · 1 SOURCE · AWS Machine Learning

Automating Amazon Textract Custom Queries adapter lifecycle across accounts

An AWS post describes architecture, infrastructure templates (CloudFormation and Terraform), and operational patterns to automate promotion, routing, and production hardening of Amazon Textract Custom Queries adapters across accounts. It shows a pre-classification routing step using DetectDocumentText and storing adapter IDs in AWS Systems Manager Parameter Store to enable zero-downtime adapter updates, plus recommended security controls (KMS, PrivateLink, least-privilege IAM, CloudTrail/CloudWatch).

5.0