At a glance概览
- LLM人工智能：大语言模型 From 0 to 1 — co-created speculator trainer for open weights models. Model customization, optimization, evaluation and benchmark.从零到一 —— 联合创建面向开放权重模型的 speculator 训练器。模型定制、优化、评估与基准测试。
- LLM人工智能：大语言模型 From 0 to 1 — co-created Anthropic inference engine: LLM inference, benchmark, computer use, Trainium, trust & safety.从零到一 —— 联合创建 Anthropic 推理引擎：大模型推理、基准测试、computer use、Trainium、信任与安全。
- Foundational Data Services基础数据服务 400TN+ objects, exabytes of data, 100M+ requests per second, 11 9's data durability.400TN+ 对象，数据规模达 EB 级，每秒 1 亿+ 请求，11 个 9 的数据持久性。
- Serverless & Virtualization无服务计算及虚拟化 From 0 to 1 — co-created AWS App Runner. UTL & TLM, overseeing two public products.从零到一 —— 联合创建 AWS App Runner。担任 UTL 与 TLM，负责两款公开产品。
- Containers容器 Incubating new service in stealth mode: MicroVM-based long-running compute platform.在隐身模式下孵化新服务：基于 MicroVM 的长时运行计算平台。
- Conversational AI人工智能：对话式 AI From 0 to 1 — co-created Health & Wellness. 100M+ users worldwide.从零到一 —— 联合创建 Health & Wellness，全球 1 亿+ 用户。
- Fintech金融科技 Scaling hundreds of billions in ARR — bill run, commerce platform.支撑数千亿美元 ARR 规模 —— 账单运行（Bill Run）、商业平台。
- Onboard入职 Joined AWS on its 10th anniversary — Pi Day, 2016.在亚马逊云科技十周年之际加入 —— 2016 年 Pi Day。
2026
- Parallelize speculative decoding with P-EAGLE on Amazon SageMaker AILLM · AI Platforms Launch blog
- Amazon SageMaker AI now supports optimized generative AI inference recommendationsLLM · AI Platforms What's newLaunch blog
- AWS 20th AnniversaryProjectversary · Amazon Matt Garman post
2025
- Kicking off NeurIPS 2025 with workshops in Mexico City!NeurIPS Kicking off NeurIPS 2025 with workshops in Mexico City!NeurIPS CDMX 2025: Accepted Workshops and TutorialsNeurIPS Announces Second Physical Location in Mexico CityReflecting on the Inaugural NeurIPS Position Paper Track: A Pilot Year JourneyNeurIPS 2025 Position Paper TrackNeurIPS 2025 Organizing Committee
- Amazon SageMaker AI now supports EAGLE speculative decodingLLM · AI Platforms What's newLaunch blogAWS re:Invent 2025 - Scale AI agents with custom models using Amazon SageMaker AI & SGLang (AIM387)Transform AI development with new Amazon SageMaker AI model customization and large-scale training capabilities by Ankur Mehrotra
- Seven Years of FirecrackerProjectversary Seven Years of Firecracker, Marc Brooker
- Token Counting Meets Amazon BedrockLLM · Bedrock Core Platform Token Counting Meets Amazon BedrockOptimizing Generative AI Workloads with Amazon Bedrock TokenizerMonitor your token usage by counting tokens before running inferenceAPI Reference CountTokens
- Amazon Bedrock launches with Claude 3.5 Sonnet in the AWS Top Secret cloudLLM · Bedrock Core Platform AWS Public Sector Blog
2024
- Big Technology | AWS CEO Matt Garman Talks Amazon’s Big Bets in AI Chips, Reasonre:Invent Big Technology | AWS CEO Matt Garman Talks Amazon’s Big Bets in AI Chips, Reasoning, and Nuclear EnergyThe Claude in Amazon Bedrock CourseLive updates from AWS re:Invent 2024AWS News Blogre:Invent RecapAWS re:Invent 2024 - Andy Jassy shares his final thoughts from re:InventAWS re:Invent 2024 - Generative AI in action: From prototype to production (AIM276)AWS re:Invent 2024 - Keynote with Dr. Swami SivasubramanianAWS re:Invent 2024 - Compute innovation for any application, anywhere (CMP215)AWS re:Invent 2024 - CEO Keynote with Matt GarmanAWS re:Invent 2024 - Monday Night Live with Peter DeSantisAWS Machine Learning Blog | Advancing AI trust with new responsible AI tools, capabilities, and resourcesAWS Machine Learning BlogContainersThe future of Kubernetes on AWS (KUB201)SiliconANGLE theCUBE | Exclusive: Amazon CEO Andy Jassy reveals AWS’ strategy for building the enterprise AI platformCMSWIRE | AWS CEO Q&A: Matt Garman on Amazon’s $4B Anthropic Bet and AI StrategySwami Sivasubramanian post December 06 2024Swami Sivasubramanian post December 05 2024Swami Sivasubramanian post December 04 2024Swami Sivasubramanian post December 03 2024Matt Garman post December 02 2024Swami Sivasubramanian post December 02 2024John Furrier post December 01 2024Matt Garman post December 01 2024SiliconANGLE theCUBE | AWS re:Invent 2024: CEO Matt Garman unveils the future of cloud with generative AI and agentic workflowsSwami Sivasubramanian post December 01 2024SiliconANGLE theCUBE | Inside the AI revolution: Swami Sivasubramanian on generative AI, agentic systems and AWS’ visionBaskar Sridharan post November 30 2024AIM276-INT
- Introducing latency-optimized inference for foundation models in Amazon BedrockLLM · Bedrock Core Platform AWS re:Invent 2024 - AWS Trainium2 for breakthrough AI training and inference performance-CMP333-NEWBusiness Insider | Amazon's AI chip executive tells BI why Nvidia is not a competitor, how Anthropic helps, and what AMD needsAWS Announces New Data Center Components to Support AI Innovation and Further Improve Energy EfficiencySemiAnalysis | Amazon’s AI Self Sufficiency | Trainium2 Architecture & NetworkingBill Jia postAWS Partners LIVE! at re:Invent 2024 | Day 1 Jason Clinton, CISO AnthropicBUSINESS WIRE | AWS Trainium2 Instances Now Generally AvailableClaude 3.5 Haiku on AWS Trainium2 and model distillation in Amazon BedrockWhat's New
- Anthropic names AWS its primary training partner and will use AWS Trainium to train and deploy its largest foundation models; Amazon to invest additional $4 billion in Anthropic.LLM · Bedrock Core Platform Anthropic NewsAmazon NewsAndy Jassy postMatt Garman post
- Celebrating 10 years of AlexaProjectversary "You and Me" | 10 Years of Alexa
- Anthropic and Palantir Partner to Bring Claude AI Models to AWS for U.S. Government Intelligence and Defense OperationsLLM · Bedrock Core Platform Palantir Investor RelationsTechCrunchBUSINESS WIRE | Anthropic and Palantir Partner to Bring Claude AI Models to AWS for U.S. Government Intelligence and Defense Operations
-
Claude 3.5 Haiku (GA) on Amazon Bedrock. (Tepui)LLM · AI/ML Services & Infrastructure Org
AWS News BlogAbout Amazon BlogWhat's NewAdam Seligman + Anthropic Partnership VideoSwami Sivasubramanian postVasi Philomin postArtificial Analysis | Claude 3.5 Haiku: API Provider Benchmarking & AnalysisWhat you need to know about the AWS AI chips powering Amazon's partnership with Anthropic4 ways AWS is engineering infrastructure to power generative AI
Internal notes (6)内部评价（6）
- SVP, L11 11/04/2024 8:28 PM
- VP, AI and Data, L10 11/04/2024 8:15 PM
- VP, L10 11/04/2024 7:16 PM
- VP, L10 11/04/2024 6:48 PM
- VP/Distinguished Engineer, L10 11/04/2024 6:36 PM
- Sr. Principal, ML Engineering, L8 11/04/2024 6:22 PM
- 10 Years of Serverless with AWS Lambda and Amazon ECSProjectversary 10 Years of Serverless with AWS Lambda and Amazon ECSServerless Vision | Move Faster from Idea to Business Value10 years of Lambda & ECS! | Serverless Office HoursServerless moments with AWS Lambda and Amazon ECSAWS Lambda: A Decade of Transformation[Interview] Holly Mesrobian, AWS VP Serverless Compute[Blog] AWS Lambda turns 10: A rare look at the doc that started it, Werner Vogels, VP & CTO at Amazon[Interview] The 10-year journey of AWS Lambda: How our serverless vision shaped the future, Werner Vogels, VP & CTO at Amazon[Blog] Ten Years of AWS Lambda, Marc Brooker, VP & Distinguished Engineer at AWSAWS News Blog | AWS Lambda turns ten – looking back and looking aheadAWS News Blog | Celebrating 10 Years of Amazon ECS: Powering a Decade of Containerized Innovation
- Upgraded Claude 3.5 Sonnet model from Anthropic introduces groundbreaking computer use capability. (Firefly)LLM · AI/ML Services & Infrastructure Org Amazon Q3 2024 Earnings ReportDeveloping a computer use modelIntroducing computer use, a new Claude 3.5 Sonnet, and Claude 3.5 HaikuAmazon Bedrock introduces an upgraded Claude 3.5 Sonnet, Anthropic’s most intelligent AI model to dateWhat's NewAWS News BlogAndy Jassy postMatt Garman postSwami Sivasubramanian postVasi Philomin postComputer use, Claude 3.5, and the race for model-to-app supremacyAWS DevOps & Developer Productivity BlogDeepak Singh postSrini Iragavarapu postClaude 3.5 Sonnet on GitHub CopilotGitHub Blog / News & insights / ProductInvestor Relations | Amazon.com Announces Third Quarter ResultsArtificial Analysis | Claude 3.5 Sonnet (Oct.): API Provider Benchmarking & AnalysisDario Amodei: Anthropic CEO on Claude, AGI & the Future of AI & Humanity | Lex Fridman Podcast #452How Amazon rebuilt Alexa with generative AIClaude and Alexa+AWS Summit New York 2024 - Keynote with Matt WoodEmpowering everyone with GenAI to rapidly build, customize, and deploy apps securely: Highlights from the AWS New York Summit
- Bedrock 1 Year Anniversary Celebration/HHProjectversary · AI/ML Services & Infrastructure Org —
- Amazon Bedrock offers select FMs for batch inference at 50% of on-demand inference priceLLM · AI/ML Services & Infrastructure Org, “Cousin” Team Engagement What's New
- Claude 3.5 Sonnet and Claude 3 Haiku now available in more regionsLLM · AI/ML Services & Infrastructure Org AWS News Blog
- Claude 3.5 Sonnet (Dove)LLM · AI/ML Services & Infrastructure Org Anthropic NewsAmazon Bedrock introduces Anthropic's Claude 3.5 Sonnet to customers, their most powerful AI model to dateAWS News Blog
- Amazon Bedrock now available in the (London), (São Paulo), and Canada (Central) regionsLLM · AI/ML Services & Infrastructure Org What's New
- Amazon 30th AnniversaryProjectversary The Economist | What next for Amazon as it turns 30?Statista | Amazon at 30: All Grown Up
- Kubernetes is turning 10!Projectversary · SILICON VALLEY, CA + VARIOUS WORLDWIDE LOCATIONS KuberTENes Birthday Bash10 Years of KubernetesWhy is Kubernetes so special? BY SARAHNOVOTNYThe Technical History of Kubernetes
- Claude 3 Sonnet and Haiku now available in Amazon Bedrock in the Europe (Frankfurt) regionLLM · AI/ML Services & Infrastructure Org What's New
- Anthropic’s Claude 3 Opus model is now available on Amazon Bedrock (Teal)LLM · AI/ML Services & Infrastructure Org Amazon Q1 2024 Earnings ReportAmazon NewsAWS News BlogAndy Jassy postAdam Selipsky postSwami Sivasubramanian postSiliconANGLEArtificial Analysis | Claude 3 Opus: API Provider Benchmarking & AnalysisAmazon Q GAAWS Summit Los Angeles 2024 - Keynote with Matt Wood
- AWS Pi Day 2024: Use your data to power generative AIProjectversary · Amazon Building Teams That Navigate Complexity by Andy WarfieldAWS News Blog
2023
- Welcome to a New Era of Building in the Cloud with Generative AI on AWSre:Invent CEO Keynote with Adam SelipskyKeynote with Dr. Swami SivasubramanianAWS Machine Learning Blog
- Dive deep on Amazon S3FDS · Foundational Data Services (FDS) Recording STG314Recording AIM250-INTRecording STG227USENIX FAST '23 Building and operating a pretty big storage system called S3Diving Deep on S3 Consistency
- RocksDB turning 10!Projectversary TweetLinkedin Post by Dhruba Borthakur, Co-founder and CTO at Rockset (acquired by OpenAI)#IndexConf 2023 - RocksDB 10 Year Anniversary Panel
- Celebrate Amazon S3’s 17th birthday at AWS Pi Day 2023Projectversary · Amazon BlogpostS3 as an Eternal Service
2022
- SVS210-Lre:Invent SVS210-LSVS404
- A close look at AWS Fargate and AWS App RunnerServerless · AWS re:Invent 2022 CON406 Recording
- Happy 5th birthday AWS Fargate!!Projectversary · Amazon Blogpost
- AWS App Runner now supports privately accessible services within Amazon VPCServerless · Developers, Events, Containers and Serverless (DECS) AWS What's NewBlogpost
- AWS App Runner launches AWS X-Ray supportServerless · AWS Elastic Containers AWS What's NewBlogpostBlogpostTweetContainers from the CouchAWS Open Source News
- AWS App Runner adds support for Amazon VPCServerless · AWS Elastic Containers AWS What's NewBlogpostBlogpostDockerCon
2021
- AWS Announces General Availability of AWS App Runner (Project Code Name: Fusion)Serverless · Containers, Linux, and App Networking Amazon Q2 2021 Earnings ReportPress ReleaseAdam Selipsky PostInfoQ NewsContainers from the Couch
2019
- Alexa at five: Looking back, looking forwardProjectversary · Amazon About Amazon Blog by Tom TaylorAmazon Science Blog by Rohit Prasad
- New ways to manage your medications at home using AlexaConversational AI · Amazon Alexa Amazon Q4 2020 Earnings ReportAmazon NewsForbesCNBCAmazon is building a 'health & wellness' team within Alexa as it aims to upend health careAmazon nabs a top Box exec in health as it goes after the medical industry
-
Management Data Store (MDB) Disaggregation in Commercial Regions (Ops Win)Commerce Platform · AWS Payments
Cost savings from MDB teardown are estimated 8.5 million/year. Reduced the variance in P99 latency of invoice creation by around 10%.
BlogpostThe New Stack | Relational to NoSQL at Enterprise Scale: Lessons from Amazon
2018
- Europe, Middle East, and Africa (EMEA) Configurable Seller of Record (CSOR)Commerce Platform · AWS Payments On 1 July 2018, Amazon Web Services EMEA SARL (“AWS Europe”), our Luxembourg-based AWS entity, started selling our services
- All AWS Services GDPR readyCommerce Platform · AWS Payments AWS Security BlogTips for Success: GDPR Lessons LearnedThe AWS Shared Responsibility Model and GDPRAWS and the General Data Protection Regulation (GDPR)AWS Announces Cloud Infrastructure Services Providers in Europe (CISPE) Membership and Compliance with First-Ever Code of Conduct for Data Protection in the Cloud
-
FinX 10x scalingCommerce Platform · AWS Payments
Over 2017, FinX improved bill run throughput by 263%. Cut down the time of invoice creation and revenue reporting by 64%.
2017
- Announcing the New AWS Secret RegionCommerce Platform · AWS Payments AWS Public Sector Blog
2016
- Announcing the AWS Canada (Central) RegionCommerce Platform · AWS Payments AWS What's New
- Stepping Onto the AWS Commerce Platform with James GreenfieldCommerce Platform · AWS Payments Stepping Onto the AWS Commerce Platform with James Greenfield
- CELEBRATING THE 10TH ANNIVERSARY OF AMAZON WEB SERVICESOnboard · Amazon Blogpost2016 Letter to Shareholders