Mona Mona

great post on the intersection of AI workloads and Open Container Initiative "The State of OCI Artifacts for AI/ML" https://lnkd.in/gbXZykMq

47

On October 19–20, 2025, AWS services in the us-east-1 (N. Virginia) region suffered a major outage after a race condition in DynamoDB’s DNS system deleted critical DNS records. This disrupted DynamoDB and caused cascading failures across EC2, NLB, Lambda, Redshift, and other services. Full recovery was completed by 2:20 PM PDT on October 20. AWS is fixing the DNS flaw and strengthening recovery, throttling, and failover mechanisms to prevent similar incidents. https://lnkd.in/gZh4A75A

1

2 Comments

Ruokai Yin, our summer intern from Yale to Microsoft Azure, worked with Xuan Zuo, @Sattwik Deb Mishra, Hokchhay T. and Preyas Shah and me. We published a paper based on his project: https://lnkd.in/eHuG3BSt Summary of paper: Finding optimal sharding is hard especially for large LLMs. A large number of devices have to be divided into TP, PP, EP dimensions. For TP, we have to decide the tensor dimension e.g. head, hidden size to shard for every op. Additionally, TP can be one-dimensional/two-dimensional etc. This makes for a large sharding space that is hard to cover exhaustively. SoTA methods use heuristics to select sharding e.g. GSPMD, Sequence Parallelism, Helix etc. There are significant time gaps, months or years, between the discovery of each of these methods. We want to fast-forward the process of discovery by 1) using a formal representation of the sharding space, and, 2) using reinforcement learning to search the sharding space.    This paper demonstrates that our method is 3.5x better than random search, and on par with hand-tuned heuristics when looking only at 1D sharding and without allowing sequence parallelism. We aim to extend the work to multi-dimensional sharding as well as sequence parallelism. Beyond that, we wish to combine the problem of sharding with fusion, tiling and collectives as well.

58

#SearchAI 11.2 is now available on the #AWS #Marketplace for easy single click deployment. Amazon Web Services (AWS) Easiest way (and fixed cost) to deploy #HybridSearch #RAG #AI #Search on #AWS https://lnkd.in/emWAPkrv

5

Amazon AWS introduces AI-Driven Development for faster, smarter coding - About Amazon India https://lnkd.in/gnzXxKMa AWS introduced AI-Driven Development Lifecycle (AI-DLC)—an openly accessible AI-native methodology designed to help organisations use AI in the software development process for building complex systems at scale with tools and frameworks such as Kiro, Amazon Q Developer, and Strands Agents. Sivasubramanian announced at the keynote that AI-DLC is available from today as a freely accessible methodology on AWS, and as a custom workflow in Kiro

1