AI Briefing
AWS Tech Blog (Korea) news
The latest AI and developer news from AWS Tech Blog (Korea), with short summaries.
Visit website →Celltrion Healthcare Achieves 70% Product Completeness in 4 Days with OMS Build Workshop Using AI-DLC and Kiro
Celltrion Healthcare conducted an OMS build workshop leveraging AWS's AI-DLC methodology and Kiro, achieving approximately 70% product completeness through a 1-day pilot and a 3-day main workshop.
Inocras Reveals Design for Optimizing Clinical Genomics Pipeline Operations on AWS HealthOmics
Inocras achieved a 72% reduction in analysis costs and a 47% improvement in processing speed by adopting AWS HealthOmics, and disclosed the design principles of the event-driven automation and GitOps deployment system that made this possible.
UneeQ Cuts Digital Human Rendering Costs by 55% Using AWS Spot Instance Orchestration
UneeQ reduced digital human rendering costs by up to 55% by combining AWS Spot instances with custom orchestration.
Production RAG Operations Strategy on Amazon Aurora PostgreSQL with pgvector 0.8.0
This article introduces practical strategies for operating production RAG workloads on Amazon Aurora PostgreSQL using pgvector 0.8.0, covering HNSW indexing, memory optimization, and observability.
AWS Accelerates Oracle-to-PostgreSQL Migration with Amazon Bedrock-Based Multi-Agent GAMMA
AWS released GAMMA, a multi-agent framework based on Amazon Bedrock, significantly reducing manual effort in Oracle migrations.
Optimizing Gemma 4 Serving with vLLM on Amazon EKS [Part 2: Throughput and Latency SLOs on a Single GPU]
This article introduces vLLM optimization strategies to maximize Gemma 4's SLO-compliant throughput through NVFP4 precision and the separation of interactive and batch pools.
Optimizing Gemma 4 Serving with vLLM on Amazon EKS: Cold Start Reduced from 428s to 226s
By leveraging Run:ai Streamer and cache optimizations for serving Gemma 4 with vLLM on Amazon EKS, cold start times were reduced from 428 seconds to 226 seconds.
SEGA Corporation: Leveraging Amazon DynamoDB and Amazon ElastiCache Serverless for Valkey to Launch 'Sonic Rumble Party' Globally with a Small Team
SEGA adopted DynamoDB and ElastiCache Serverless to operate 'Sonic Rumble Party' globally with a small team, significantly reducing costs.
GS Retail's Enterprise AI Gateway Operations: Budget Routing and Custom Guardrails
GS Retail shared its operations and governance case study for an enterprise AI Gateway built in collaboration with AWS.
GS Retail Builds Enterprise-Wide AI Gateway in Collaboration with AWS, Showcasing Automation of Authentication, Routing, and Account Management
GS Retail has revealed its collaboration with AWS to build an enterprise-wide AI Gateway based on LiteLLM, automating authentication, routing, and account management.
Improving OpenSearch Search Quality for AI Agents (Part 2)
Applying Rerank improves BM25 but degrades performance by up to 9% in hybrid search.
Channel Corporation Reduces Migration Time by 96% via DynamoDB Table Separation and AWS Managed Services
Channel Corporation separated Badge data into a dedicated table and utilized AWS managed services to reduce migration time by 96%.
LG Energy Solution Builds ERCOT Power Bidding Optimization Agent Based on Amazon Bedrock AgentCore
LG Energy Solution has built an ERCOT power market bidding optimization agent using Amazon Bedrock AgentCore.
Amazon ECS Execution Architecture and Selection Criteria – Part 1: Compute and Execution Types
Amazon ECS recommends using capacity providers for actual execution while treating launch types as declarative only, and introduces Managed Instances to simultaneously address GPU support and OS patch management.
Amazon ECS Execution Architecture and Selection Criteria – Part 2: Deployment Strategies, Networking, and Design Limits
Amazon ECS deployment strategies have expanded to six types, with detailed design criteria and quota limits for networking and storage options provided.
Samsung Account AIOps: Autonomous Operations Based on AgentCore for Incident Response and Report Automation
Samsung Account automated incident analysis and infrastructure change impact assessment using AgentCore-based AIOps, improving operational efficiency.
Samsung Account Builds Multi-Agent AIOps System on Amazon Bedrock AgentCore
Samsung Account built a multi-agent AIOps system using Amazon Bedrock AgentCore to address operational complexity in multi-region environments.
AWS Reduces Nitro V6 Instance TCP Idle Timeout from 5 Days to 350 Seconds
AWS has reduced the default TCP established idle timeout for Nitro V6-based instances from 5 days to 350 seconds to prevent connection errors.
Amazon Quick Supports FinOps Workflow Automation via CUDOS Dashboards
Amazon Quick integrates with CUDOS dashboards to enhance FinOps efficiency through natural language analysis and automation of repetitive tasks.
Managing Claude Code Organization Costs: Cache TTL Optimization and Amazon Bedrock Attribution Strategies
This article presents methods for managing Claude Code prompt cache TTL and controlling organizational costs in cloud environments such as Amazon Bedrock.