SEP 5, 2026
EngBrief
Search⌘K
LatestTopicsSourcesSaved
Eng&Brief

Engineering insights from the world's best tech companies, curated and summarized.

Weekly brief

Browse

TopicsSourcesFavorites

More

SearchRSS Feed
© 2026 EngBriefUpdated every 4 hours
← Sources
aws.amazon.com icon
Cloud Architecture

AWS Architecture Blog

83 articles on EngBrief

The AWS Architecture Blog provides guidance on building well-architected solutions on Amazon Web Services. Posts cover cloud design patterns, reference architectures, cost optimization strategies, and best practices for reliability, security, and performance across AWS services.

Cloud ArchitectureServerlessCost OptimizationReliability
Visit blog →

Latest Articles

AWS3d ago

Hybrid cloud orchestration: Modernizing on-premises infrastructure management with AWS

Learn how to build a hybrid cloud orchestration solution that manages distributed on-premises infrastructure at scale using AWS serverless technologies and...

CloudArchitecture
1 min
AWS3d ago

MCP went stateless: Is your AWS MCP server deployment well-architected?

On July 28, 2026, MCP made its protocol core stateless, removing the initialize handshake and session header. This post maps the MCP 2026-07-28 specification...

CloudArchitecture
1 min
AWS9d ago

Gallup scales real-time coaching for thousands with Amazon Bedrock

Gallup transformed 90 years of workplace science into Gallup AI, a generative AI assistant powered by Amazon Bedrock that delivers real-time, personalized...

CloudArchitecture
1 min
AWS9d ago

Closing the AI agent trust gap with graduated autonomy

Most teams give AI agents either full access or read-only, leaving value unused or risk unmanaged. This post describes graduated autonomy, an architectural...

CloudArchitecture
1 min
AWS11d ago

How a global payment processor preserved AWS RAM shares and Lake Formation permissions during an AWS Organizations migration

When AWS accounts move between organizations, organization-bound AWS RAM resource shares break and control-plane access is lost. Learn how a global payment...

CloudArchitecture
1 min
AWS14d ago

Build a unified AI agent architecture with DynamoDB and Bedrock

With native vector search in Amazon DynamoDB, you can store vector embeddings alongside your operational data in a single table. This post shows how to build a...

CloudArchitecture
1 min
AWS14d ago

How AgentFlo built AI sales agents with Amazon Bedrock AgentCore – Part 2

Part 2: how AgentFlo built trusted, reliable AI sales agents on Amazon Bedrock AgentCore and AWS serverless architecture. Learn the three-layer guardrails,...

CloudArchitecture
1 min
AWS16d ago

How AgentFlo built AI sales agents with Amazon Bedrock AgentCore – Part 1

Learn how AgentFlo built always-on AI sales agents on Amazon Bedrock AgentCore and the Strands Agents SDK. Part 1 covers three pillars of production-grade...

CloudArchitecture
1 min
AWS16d ago

How Clario technology detects PHI/PII in DICOM images using Amazon Bedrock

Clario, part of Thermo Fisher Scientific, uses Amazon Bedrock and Amazon Textract to automatically detect protected health information (PHI) and personally...

CloudArchitecture
1 min
AWS16d ago

AI-powered clinical trial eligibility and safety using Amazon Bedrock AgentCore

AI agents built on Amazon Bedrock AgentCore help clinical trial teams make fast, accurate enrollment decisions while keeping clinicians in control. This post...

CloudArchitecture
1 min
AWS17d ago

Consistency is the new latency: AI at the data layer

As AI agents move from chatbots to taking action, their reliability depends on the consistency of the data layer beneath them. This post examines how...

CloudArchitecture
1 min
AWS21d ago

Serverless vehicle tracking at scale: Bosch L.OS on AWS

Bosch Mobility Platform Solutions developed L.OS, a serverless platform on AWS that standardizes vehicle tracking across India's fragmented spot logistics market. L.OS integrates multiple tracking data sources and provides real-time visibility and data-driven decision-making across complex logistics networks. The platform solves key challenges, including data fragmentation, incompatible data formats, and variable data quality, enabling scalable and efficient logistics operations. L.OS uses AWS services, including Amazon ECS Fargate, AWS Lambda adapters, Amazon MSK, and Amazon QuickSight, to create a scalable and maintainable system. The architecture ensures secure and real-time communication between clients, tracking providers, and the L.OS platform, handling thousands of concurrent tracking requests and providing accurate ETAs. The key components of the architecture include centralized orchestration with Amazon ECS Fargate, serverless provider integration using AWS Lambda, event-driven communication with Amazon MSK, and subscription and access management with AWS Marketplace Subscription Management service. These components enable seamless protocol translation, event streaming,

CloudArchitecture
1 min
AWS22d ago

Track generative AI costs with Amazon Bedrock inference profiles

Here's a 2-3 sentence summary of the blog post on tracking generative AI costs with Amazon Bedrock inference profiles: To track costs for multiple teams using a shared Amazon Bedrock foundation model, create an application inference profile for each team with a unique cost allocation tag and associate it with the same model. Route application invocations through department-specific profiles and use AWS cost allocation tags to view per-department Amazon Bedrock costs as separate line items in AWS Cost Explorer. This solution can track costs for each team without adding per-user session management and pays the same per-token rate as invoking the model directly, with no additional charges for cost attribution.

CloudArchitecture
1 min
AWS22d ago

Recovery strategies to meet data residency requirements

Here's a 2-3 sentence summary for each of the recovery strategies: **Strategy 1: Cryptographic Boundary** This strategy uses encryption and AWS Key Management Service (AWS KMS) keys to replicate data into another AWS Region in a geopolitically aligned country, ensuring data protection and preventing unauthorized access. Key policies can be used to explicitly deny decryption operations in the recovery Region, giving strong assurance that data remains secure. Suitable for customers who accept storing encrypted data in a partner country and prioritize operational simplicity. **Strategy 2: Data Boundary** In this approach, customers store backups or operate a pilot light recovery environment on AWS Outposts in an on-premises site within the source country or an approved location, maintaining full control of business data physically residing at all times. Replication is done using tools such as AWS DataSync or third-party backup solutions to replicate data from the primary Region to on-premises or in-country storage. This strategy offers consistency in using AWS services and hardware

CloudArchitecture
1 min
AWS23d ago

Reducing Text2SQL latency with parameterized query templates

Here is a summary of the engineering blog post in 2-3 concise sentences: By implementing parameterized query templates with caching, a Text2SQL system achieved an 80% reduction in end-to-end latency and a 50% decrease in token consumption, turning a slow prototype into a responsive production system. The solution uses a template caching layer where similar requests can share cached results, leveraging vector embeddings for semantic similarity search and named entity recognition for filling template placeholders. This approach improves scalability, reduces latency, and enables the system to preserve accuracy even under real traffic.

CloudArchitecture
1 min
AWS23d ago

Adobe Firefly: Simplified observability with Amazon Managed Prometheus

Adobe Firefly achieved significant improvements in query performance, reliability, and scale by migrating its observability architecture from a self-managed Prometheus deployment to Amazon Managed Service for Prometheus. The new architecture enabled 28x faster query performance and extended observability windows from 6 hours to 24 hours. Amazon Managed Service for Prometheus handled the collection of metrics from Adobe's Amazon EKS-based training clusters, providing high availability, scalability, and ease of migration. The team at Adobe Firefly now views metrics across 24-hour windows, making it easier to identify performance degradation and make informed decisions. Additionally, the reduced operational overhead of Amazon Managed Service for Prometheus allowed the team to minimize the burden of maintaining self-hosted Prometheus infrastructure. The migration to Amazon Managed Service for Prometheus also laid the groundwork for future scalability and high availability, with the ability to handle up to 50 million active time series and support for up to 1 billion.

CloudArchitecture
1 min
AWS24d ago

Scaling patterns for self-organizing multi-agent clusters with Kiro

Kiro CLI's self-organizing multi-agent clusters rely on a supervisor agent breaking tasks into pieces, which is replaced by a shared state in this scaling pattern. The pattern involves independent agents reading and writing to a shared store, with each agent contributing to a goal without a central planner or gate. This approach allows for more parallelism and diversity in approaches, reducing errors and increasing productivity, particularly when workloads involve many quasi-independent tasks or contributions. However, it requires a known task tree with strict ordering for a supervisor to ensure central verification, and may not provide a gate between each step.

CloudArchitecture
1 min
AWSJul 22, 2026

Building a serverless AI assistant at Pelago: concept to care in two weeks

The Pelago engineering team built a serverless AI assistant in two weeks using AWS services such as Amazon Bedrock and AWS Lambda, enabling contextually aware suggested considerations for the care team. The event-driven architecture separates concerns, allowing the AI assistant to process messages asynchronously without blocking the user experience. The system generates pre-generated suggestions, stored in a database, and retrieves them instantly when a care team member opens a conversation. The team used Amazon SNS for message fan-out and decoupled processing, adding the AI Chat Assistant feature without modifying existing message-handling code. This architecture also enabled organic scaling with each Lambda function scaling horizontally based on current traffic.

CloudArchitecture
1 min
AWSJul 22, 2026

Building multi-Region resiliency for AWS CloudFormation custom resource deployment

Here is a concise 3-sentence summary of the engineering blog post: A team at AWS designed an active-active multi-Region architecture to provide resiliency for AWS CloudFormation custom resource deployment, addressing native gaps in fan-out, execution, and failover mechanisms. The solution utilizes Amazon DynamoDB Global Tables for distributed locking, idempotency, and state tracking, along with Amazon Application Recovery Controller for automated failover and Amazon SNS for cross-Region event fan-out. This architecture enables mission-critical workloads to handle regional failures, achieving high availability and preventing duplicate processing risks in multi-Region custom resource designs.

CloudArchitecture
1 min
AWSJul 22, 2026

Architecting offline-first generative AI applications for edge deployments using AWS services

Here's a 3-sentence summary of the blog post: A reference architecture for offline-first generative AI applications at the edge is presented, using AWS services including Amazon Bedrock, SageMaker AI, AWS IoT Greengrass, and Strands Agents. The architecture addresses the challenge of customizing large-scale AI models for reliability and scalability, using a hybrid approach that combines fine-tuning, continued pre-training, and retrieval-augmented generation (RAG) to optimize performance and accuracy. The design involves careful coordination across multiple AWS services and deliberate trade-offs between model capability, hardware constraints, and operational complexity to achieve a cost-effective and scalable edge-deployment of generative AI applications.

CloudArchitecture
1 min