Skip to content
Thursday, 19 November 2026
Tenesys AI News
Subscribe

AI Automation

Workflow automation, AI customer support, contact-center and CRM automation.

“Amazon” — 4 articles ✕

Automation· Important· 🧪 Worth Testing

Detecting and preventing silent data corruption in large-scale GPU training

Amazon introduced a framework for identifying silent data corruption (SDC) in distributed training, where faulty hardware returns incorrect math without triggering standard error logs. By integrating three custom checks—gradient sanity checks, ring replay audits, and weight fingerprinting—into PyTorch DDP training loops, you can identify, isolate, and replace failing nodes.

AWS Machine Learning
Automation· Important· 🧪 Worth Testing

Amazon SageMaker now supports Multi-Instance GPU for managed endpoints

Amazon SageMaker Inference has introduced support for NVIDIA Multi-Instance GPU (MIG), allowing you to partition a single physical GPU into multiple independent units. This hardware-level division provides dedicated compute and memory to each partition, ensuring predictable latency and strong isolation between models hosted on the same accelerator.

AWS Machine Learning
Automation· Important· 🧪 Worth Testing

AWS launches optimized Ray Train Deep Learning Containers for EKS

AWS introduced pre-built Ray Train Deep Learning Containers (DLCs) that consolidate PyTorch, CUDA, NCCL, and EFA networking libraries into a single tested image. This approach allows your team to deploy distributed training clusters on Amazon EKS without maintaining custom build pipelines or resolving complex driver compatibility issues.

AWS Machine Learning
Automation· Important· 🧪 Worth Testing

Scaling agentic QA automation with multi-team governance

AWS has updated its QA Studio reference solution, adding native support for multi-team organizational structures. The new capabilities include application-level ownership, automated health scoring based on CloudWatch metrics, and granular role-based access control.

AWS Machine Learning