Bedrock Brief 08 Oct 2025

Pleri AI

Pleri AI

— 3 min read
Bedrock Brief 08 Oct 2025

Welcome to this week's Bedrock Brief, where we're diving deep into the AI waters and coming up for air with some juicy AWS insights.

First up, it seems the race for AI talent is hotter than a overclocked GPU. An AWS applied scientist is making waves, claiming that PhDs in AI will become even more valuable as the tech evolves. Sure, you could learn on the job, but if you want to shape the future of AI, apparently you need those fancy letters after your name. Who knew that spending five years eating nothing but stew could be your ticket to a million-dollar payday?

Meanwhile, in a move that's equal parts exciting and terrifying, Commonwealth Bank of Australia has decided to go all-in on AWS, migrating their entire core banking system to the cloud. Picture this: your savings account balance now floats somewhere in the digital ether, powered by Amazon's data centers. It's like your money is taking a permanent vacation in the cloud – let's just hope it doesn't get lost in the baggage claim.

Lastly, AWS is throwing a global AI coding party and you're all invited. The AWS Global Vibe Hackathon 2025 is dangling over $700K in prizes for developers willing to flex their AI muscles. From agentic systems to AI-powered education, the tracks are as diverse as a machine learning dataset. So dust off your coding fingers, caffeinate accordingly, and get ready to build the future – or at least a really cool chatbot that can finally understand your sarcasm.

Fresh Cut

  • Amazon's new EC2 C8i and C8i-flex instances offer up to 15% better price-performance and 2.5x more memory bandwidth than previous Intel-based instances, making them ideal for compute-intensive workloads like web servers, databases, and AI deep learning. Read announcement →
  • Amazon OpenSearch Ingestion now enables batch AI inference for large-scale data enrichment, allowing developers to efficiently process billions of records with AI-powered features like vector embeddings and translations. Read announcement →
  • Amazon Connect's new AI feature summarizes email conversations, suggests actions, and generates responses, helping customer service agents handle emails faster and more consistently. Read announcement →
  • Amazon Neptune now works with GraphStorm, letting developers quickly apply machine learning to graph data for real-time tasks like fraud detection and personalized recommendations. Read announcement →
  • Developers can now use natural language to create and deploy AI agents for Amazon Bedrock AgentCore with the open-source Model Context Protocol server, streamlining the process from development to production. Read announcement →
  • Amazon Connect users can now collect customer input on outbound calls before connecting to an agent, enabling features like obtaining consent for call recording. Read announcement →
  • AWS Clean Rooms introduces data access budgets, allowing you to set limits on how often your shared data can be analyzed, giving you more control over data privacy in collaborative projects. Read announcement →
  • Cohere's Embed v4 model on Amazon Bedrock can now process text, images, and complex business documents in over 100 languages, making it easier for developers to build powerful search and retrieval applications without extensive data preprocessing. Read announcement →
  • Amazon Bedrock Data Automation can now distinguish between speakers in audio transcripts, making it easier for developers to analyze conversations and extract insights from multi-party voice recordings. Read announcement →
  • Amazon Detective users can now securely access its API from within their VPC without internet exposure, enhancing privacy and security for threat investigations and analysis. Read announcement →

The Quarry

Responsible AI: How PowerSchool safeguards millions of students with AI-powered content filtering using Amazon SageMaker AI

PowerSchool's innovative AI content filtering solution, built on Amazon SageMaker, is keeping millions of students safe online without unnecessarily restricting their digital learning. By fine-tuning the Llama 3.1 8B model, they've created a system that outperforms off-the-shelf alternatives, boasting improved accuracy and low false positive rates. The secret sauce? A clever deployment architecture that combines the power of large language models with the flexibility of custom logic, allowing for nuanced content analysis that adapts to educational contexts. Read blog →

More posts:

Core Sample

AWS for Software & Technology, Spotlight Interview, Resolve AI

Resolve AI leverages AWS to build an AI-powered platform that streamlines IT operations, tackling the challenge of managing complex IT environments. Their solution uses natural language processing to understand and respond to IT tickets, automating up to 80% of routine tasks. By tapping into AWS's scalable infrastructure and AI services, Resolve AI can rapidly iterate on their product, staying ahead of market demands in the fast-paced world of enterprise IT management. Watch video →

More videos:

Read more