Linked and quoted from the labs, the press and the people using them. Times in UTC.
Every development we picked up, including smaller stories beyond the labs. Newest first, times in UTC.
Trane Technologies built an AI agentic solution on Amazon Bedrock AgentCore in about four weeks. It reduced a 20-minute multi-screen building diagnostic workflow to a 20-second natural language interaction.
AWS published a post on combining Amazon Bedrock Data Automation with the Model Context Protocol to structure unstructured public sector evidence. The structured insights are surfaced through natural language queries in Salesforce Agentforce.
Posit's data science IDE Positron now runs on Amazon SageMaker AI. The post demonstrates exploring an Athena table, validating features in R, training an XGBoost model in Python, deploying a real-time SageMaker AI endpoint, and reporting with Quarto in a SageMaker Studio Space.
Benchling built a defense-in-depth security architecture to run untrusted AI agent-generated scientific code across thousands of life sciences tenants using Amazon Bedrock AgentCore Code Interpreter in VPC mode. It combines Amazon Route 53 Resolver DNS Firewall and VPC endpoint policies to block data exfiltration,…
EXL built a medical intelligent document processing solution on AWS using domain-specific large language models on Amazon SageMaker and Amazon Bedrock. The system extracts, summarizes, and queries medical records at enterprise scale and cuts claims review time from over 100 minutes per case.
Amazon Web Services launched Strands Harness, an open-source AI agent for building and deploying AI applications in any environment. AWS said it aims to help developers scale AI agents to cloud environments.
Amazon announced a new AgentCore runtime, a capability of Amazon Bedrock AgentCore. It reclaims memory as sessions release it and delivers consistent cold starts regardless of image size or concurrency.
Amazon SageMaker AI now supports deploying production-ready Hugging Face models using six open-source agent skills. A coding agent can produce a real-time endpoint with the right serving container, autoscaling, CloudWatch alarms, and a verified teardown path.
Amazon introduced SageMaker HyperPod Inference Gateway, a Kubernetes-native GPU-aware routing add-on for Amazon EKS. It routes inference requests using real-time GPU signals and cuts first-token latency by up to 82% without changing model servers or clients.
Amazon announced Amazon Connect Talent, an AI hiring solution for talent acquisition teams managing scaled hiring. It offers AI-led interviews, data-driven assessments and consistent evaluation, with recruiters retaining final hiring decisions.
Amazon will use Twitch streamers' content to train its AI by default unless creators opt out. Twitch CPO Mike Minton said on a livestream that an opt-in approach would result in nobody opting in.