Latest publications
How TGS and AWS Reduced Seismic AI Model Training from Six Months to Five Days
TGS and AWS achieved nearly linear scaling of seismic foundation-model training, reducing the cycle from six months to five days and increasing the context window by 4.5x.
Guidesly shows how Jack AI on AWS transforms travel media into reports and marketing
Guidesly built a pipeline on AWS for Jack AI: the service takes photos and videos from trips, enriches them with context, analyzes them through CV and LLM, and produces ready-made reports and sales content.
AWS shows how speculative decoding on Trainium2 accelerates generation in vLLM
AWS explained how speculative decoding on Trainium2 and vLLM helps speed up generation in LLM services, reduce token cost, and increase throughput on long-response workloads.
Amazon Bedrock Added Formal Verification of AI Responses for Compliance Tasks
AWS is advancing Automated Reasoning checks in Amazon Bedrock: the service formally proves the correctness of model responses against specified rules and simplifies auditing in regulated industries.
AWS Presented a Practical Guide to Fine-Tuning Amazon Nova via Nova Forge SDK
AWS released a practical guide on fine-tuning Amazon Nova in Nova Forge SDK: from data preparation and blending to model training and result verification.
AWS shows how to cut semantic video search costs with Amazon Nova on Bedrock
AWS demonstrated how distillation in Amazon Bedrock transfers routing logic from Nova Premier to Nova Micro, reducing inference cost by more than 95% and latency by 50%.
AWS Shows How to Build AI Agents on SageMaker and Test Models via MLflow
AWS released a guide on building AI agents using Strands Agents SDK with models in SageMaker, tracing in MLflow, and working A/B tests between model variants.
Popsa Automated Personalized Titles Across 12 Languages with Amazon Nova
Popsa updated title generation using Amazon Bedrock, Nova, and Claude 3 Haiku: the system works in 12 languages, responds faster, and has already generated over 5.5 million personalized titles in 2025.
AWS showed Amazon Quick Flows — AI automation of tasks without code for business
AWS showed Amazon Quick Flows — a service that builds AI processes without code based on text description: from real-time financial analytics to employee onboarding.
AWS Showed How to Automatically Synchronize Amazon Bedrock Knowledge Bases via S3
AWS described a serverless architecture that tracks changes in S3, automatically launches ingestion jobs for Amazon Bedrock Knowledge Bases, manages API quotas, and provides monitoring.
Amazon Bedrock AgentCore and Nova 2 Sonic: Building Omnichannel Orders on AI Agents
AWS showed how to build a complete omnichannel order system based on AgentCore and the Nova 2 Sonic voice model—without being tied to a specific framework.
AWS Releases ToolSimulator for Safe Testing of AI Agents in Strands Evals
AWS has opened access to ToolSimulator — an LLM framework within the Strands Evals SDK for large-scale testing of AI agents without real API calls and the risk of data leakage.
Amazon SageMaker and DVC: end-to-end traceability of ML models from data to prediction
AWS showed how to combine DVC and SageMaker MLflow Apps for full traceability of the ML pipeline — at the dataset and individual-record levels.
CI/CD Pipeline for Amazon Lex: Conflict-Free Team Development
How isolated development environments and automated testing help teams scale chatbots on Amazon Lex without chaos.
AWS teaches Strands Agents to work with any model on SageMaker
Amazon showed how to connect language models deployed on SageMaker to the Strands Agents framework, even if they do not support the standard Bedrock API. We explain why this changes the game for enterprise AI.
AWS updated its container for running large language models: what changed and why it matters
Amazon Web Services released a major update to the Large Model Inference container, promising a noticeable performance boost and simpler LLM deployment. Here's what that means for the industry.
Amazon Bedrock AgentCore: smart assistant for events
AWS showed how to deploy a production-ready AI agent with memory and personalization in a matter of hours — without writing custom infrastructure.
How language models help train construction robots without manual data labeling
Bedrock Robotics and AWS have found a way to scale data preparation for autonomous construction machinery — instead of an army of annotators, they use vision-language models that analyze videos from construction sites.
Hugging Face and AWS build an open-source medical AI agent
Hugging Face's smolagents library now integrates with AWS cloud services to build agentic systems. The first use case is a clinical AI assistant with support for multiple models and search across medical knowledge bases.
Amazon QuickSight introduces secure authentication for Snowflake
Amazon QuickSight has added support for key-pair authentication for Snowflake, providing a new level of security for cloud analytics and enterprise data.
Amazon Bedrock AgentCore: building unified intelligent systems for business
Amazon introduced AgentCore, a new tool in the Bedrock ecosystem for bringing AI agents and knowledge bases together into a unified intelligent network.
Claude 4.5 and Claude Code come to AWS's secure cloud
Anthropic and Amazon are expanding options for the public sector. Claude 4.5 and the Claude Code agent are now available in AWS's secure environment for working with regulated data.
NVIDIA Nemotron 3 Nano 30B MoE is now available in Amazon SageMaker
NVIDIA is expanding its presence in the AWS cloud: the compact Nemotron 3 Nano 30B MoE model is now available to SageMaker JumpStart users for rapid deployment of AI solutions.
Amazon Nova Sonic: a new standard for real-time voice assistants
Amazon introduces Nova Sonic, a technology that replaces complex cascaded systems with a single interface for natural communication with AI.
Iberdrola deploys Amazon Bedrock AI agents to optimize IT processes
Energy giant Iberdrola is shifting IT infrastructure management toward generative AI. With Amazon Bedrock, the company is speeding up incident handling and automating request review.
Amazon speeds up new warehouse launches with Nova AI models
Amazon is deploying multimodal Nova models to automate equipment inspections in logistics centers, replacing manual checks with computer vision.
Neural network trained to check and rewrite code
# Neural Network Trained to Check and Rewrite Code Machine learning has reached one of the most tedious and critical tasks for developers — code review and improvement.
Marketing Image Generation with AI: Learning from Past Experience
## AI-Powered Marketing Image Generation: Learning from Past Experience Continuing the theme of using generative AI for marketing, Amazon presents a solution for creating images based on the…
Amazon Nova: now your data understand each other without words and tags
Remember the times when searching a video archive turned into endless timeline scrolling or hoping that some intern tagged things correctly? Amazon decided it's time to end this.
AWS and Google Drive: How to Build a Bridge Between Feuding Clouds
Imagine living in a world where two neighbors in a stairwell haven't greeted each other for years out of principle. One neighbor is Amazon, the other is Google.