56,966 blogs · [ { "id": "01a0e37b-ba3b-71db-b11b-2648bde55b59", "title": "A kernel-centric path to real-time video generation on Trainium", "url": "https://www.amazon.science/blog/a-kernel-centric-path-to-real-time-video-generation-on-trainium", "published_at": "2026-09-25T14:55:00+00:00" }, { "id": "01a0d526-24ad-739f-9039-1ca87f9200f9", "title": "Amazon launches research initiative with Stanford University to advance AI and science", "url": "https://www.amazon.science/news/amazon-launches-research-initiative-with-stanford-university-to-advance-ai-and-science", "published_at": "2026-09-21T19:31:59+00:00" }, { "id": "01a0d526-24ad-739f-9039-1ca8801bc62a", "title": "Advancing AI for biology: Teaching models to design and characterize antibodies", "url": "https://www.amazon.science/blog/advancing-ai-for-biology-teaching-models-to-design-and-characterize-antibodies", "published_at": "2026-09-21T17:51:00+00:00" }, { "id": "01a08c6f-0ba2-73c9-9e80-4f0055dac4a2", "title": "Why don’t machine learning research agents overfit?", "url": "https://www.amazon.science/blog/why-dont-machine-learning-research-agents-overfit", "published_at": "2026-09-10T15:03:39+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbaec792b8", "title": "Developing provably correct Rust code with Verus", "url": "https://www.amazon.science/blog/developing-provably-correct-rust-code-with-verus", "published_at": "2026-08-31T15:35:33+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbafac82da", "title": "When LLM judges agree, should we believe them?", "url": "https://www.amazon.science/blog/when-llm-judges-agree-should-we-believe-them", "published_at": "2026-08-26T17:10:40+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb0381976", "title": "SOP-Bench: A new benchmark for evaluating AI agents on real business procedures", "url": "https://www.amazon.science/blog/sop-bench-a-new-benchmark-for-evaluating-ai-agents-on-real-business-procedures", "published_at": "2026-08-21T15:57:17+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb09fb898", "title": "A decade of mathematical certainty: Reflections on the Automated Reasoning Group", "url": "https://www.amazon.science/blog/a-decade-of-mathematical-certainty-reflections-on-the-automated-reasoning-group", "published_at": "2026-08-11T16:22:19+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb15bb776", "title": "AWS Trainium Frontier competition: Co-design models and kernels on purpose-built AI chips", "url": "https://www.amazon.science/news/aws-trainium-frontier-competition-co-design-models-and-kernels-on-purpose-built-ai-chips", "published_at": "2026-08-10T20:23:04+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb24b89d7", "title": "34 Amazon Research Awards Build on Trainium recipients announced", "url": "https://www.amazon.science/research-awards/latest-news/34-amazon-research-awards-build-on-trainium-recipients-announced", "published_at": "2026-08-05T15:00:00+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb32c193a", "title": "How controllers from industrial machinery can coordinate multitask machine learning", "url": "https://www.amazon.science/blog/how-controllers-from-industrial-machinery-can-coordinate-multitask-machine-learning", "published_at": "2026-07-30T17:26:47+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb38838ee", "title": "A new benchmark for evaluating patient-facing health AI agents", "url": "https://www.amazon.science/blog/a-new-benchmark-for-evaluating-patient-facing-health-ai-agents", "published_at": "2026-07-29T15:16:52+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb3f246af", "title": "Amazon is investing in the Lean Focused Research Organization", "url": "https://www.amazon.science/news/amazon-is-investing-in-the-lean-focused-research-organization", "published_at": "2026-07-26T08:00:00+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb4e7074f", "title": "Amazon and University of Michigan give robots a sense of touch", "url": "https://www.amazon.science/blog/amazon-and-university-of-michigan-give-robots-a-sense-of-touch", "published_at": "2026-07-10T17:13:31+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb5918dbf", "title": "Capturing token IDs during agentic interactions for better reinforcement learning", "url": "https://www.amazon.science/blog/capturing-token-ids-during-agentic-interactions-for-better-reinforcement-learning", "published_at": "2026-07-09T12:46:00+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb5f375b7", "title": "How Amazon tracks carbon intensity across its operations", "url": "https://www.amazon.science/news/how-amazon-tracks-carbon-intensity-across-its-operations", "published_at": "2026-07-01T15:56:42+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb66e69d0", "title": "The fuel of the future is already here: Why TRISO matters", "url": "https://www.amazon.science/blog/the-fuel-of-the-future-is-already-here-why-triso-matters", "published_at": "2026-06-24T19:57:09+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb757c5e6", "title": "EC2’s formally verified “isolation engine” provides mathematical assurance of virtual-machine isolation", "url": "https://www.amazon.science/blog/ec2s-formally-verified-isolation-engine-provides-mathematical-assurance-of-virtual-machine-isolation", "published_at": "2026-06-10T15:00:00+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb8100fc4", "title": "Graviton5’s improved design increases speed and energy efficiency — beyond Moore’s law", "url": "https://www.amazon.science/blog/graviton5s-improved-design-increases-speed-and-energy-efficiency-beyond-moores-law", "published_at": "2026-06-10T15:00:00+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb829a9c8", "title": "Real-world grounding in agentic AI", "url": "https://www.amazon.science/blog/real-world-grounding-in-agentic-ai", "published_at": "2026-06-08T19:00:00+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb84f1a6d", "title": "Bridging intent and execution in agentic systems", "url": "https://www.amazon.science/blog/bridging-intent-and-execution-in-agentic-systems", "published_at": "2026-06-08T17:00:00+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb8cbd733", "title": "Ground truth is a process, not a dataset", "url": "https://www.amazon.science/blog/ground-truth-is-a-process-not-a-dataset", "published_at": "2026-06-03T15:56:57+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb95ce020", "title": "How flat is replacing fat in AWS data center networks", "url": "https://www.amazon.science/blog/how-flat-is-replacing-fat-in-aws-data-center-networks", "published_at": "2026-05-28T10:30:00+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbb9d6a85d", "title": "Amazon Research Awards recipients announced", "url": "https://www.amazon.science/research-awards/latest-news/fall-2025-amazon-research-awards-recipients-announced", "published_at": "2026-05-27T17:21:51+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbbad35092", "title": "Diverse reasoning traces teach LLMs to make better decisions", "url": "https://www.amazon.science/blog/diverse-reasoning-traces-teach-llms-to-make-better-decisions", "published_at": "2026-05-26T15:17:06+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbbb0d5647", "title": "Making LLMs faster without sacrificing accuracy", "url": "https://www.amazon.science/blog/making-llms-faster-without-sacrificing-accuracy", "published_at": "2026-05-15T13:00:00+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbbbfa8474", "title": "Promptimus: Improving already good LLM prompts with zero manual engineering", "url": "https://www.amazon.science/blog/promptimus-improving-already-good-llm-prompts-with-zero-manual-engineering", "published_at": "2026-05-14T13:47:45+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbbc6a7ea2", "title": "Navigating uncertainty in Amazon's middle-mile network", "url": "https://www.amazon.science/blog/navigating-uncertainty-in-amazons-middle-mile-network", "published_at": "2026-05-06T13:37:38+00:00" }, { "id": "01a07dab-c992-73a9-92fc-3fcbbd18896b", "title": "How mechanism design theory helps optimize Amazon-vendor collaboration", "url": "https://www.amazon.science/blog/how-mechanism-design-theory-helps-optimize-amazon-vendor-collaboration", "published_at": "2026-05-05T13:11:23+00:00" } ] posts Claim your blog

Back to Amazon Science

Blog · corpus.blog/blogs/amazon.science/posts

Amazon Science

amazon.science

2026

A kernel-centric path to real-time video generation on Trainium original ↗

25 Sept 2026

Amazon launches research initiative with Stanford University to advance AI and science original ↗

21 Sept 2026

Advancing AI for biology: Teaching models to design and characterize antibodies original ↗

21 Sept 2026

Why don’t machine learning research agents overfit?original ↗

10 Sept 2026

Developing provably correct Rust code with Verus original ↗

31 Aug 2026

When LLM judges agree, should we believe them?original ↗

26 Aug 2026

SOP-Bench: A new benchmark for evaluating AI agents on real business procedures original ↗

21 Aug 2026

A decade of mathematical certainty: Reflections on the Automated Reasoning Group original ↗

11 Aug 2026

AWS Trainium Frontier competition: Co-design models and kernels on purpose-built AI chips original ↗

10 Aug 2026

34 Amazon Research Awards Build on Trainium recipients announced original ↗

5 Aug 2026

How controllers from industrial machinery can coordinate multitask machine learning original ↗

30 Jul 2026

A new benchmark for evaluating patient-facing health AI agents original ↗

29 Jul 2026

Amazon is investing in the Lean Focused Research Organization original ↗

26 Jul 2026

Amazon and University of Michigan give robots a sense of touch original ↗

10 Jul 2026

Capturing token IDs during agentic interactions for better reinforcement learning original ↗

9 Jul 2026

How Amazon tracks carbon intensity across its operations original ↗

1 Jul 2026

The fuel of the future is already here: Why TRISO matters original ↗

24 Jun 2026

EC2’s formally verified “isolation engine” provides mathematical assurance of virtual-machine isolation original ↗

10 Jun 2026

Graviton5’s improved design increases speed and energy efficiency — beyond Moore’s law original ↗

10 Jun 2026

Real-world grounding in agentic AI original ↗

8 Jun 2026

Bridging intent and execution in agentic systems original ↗

8 Jun 2026

Ground truth is a process, not a dataset original ↗

3 Jun 2026

How flat is replacing fat in AWS data center networks original ↗

28 May 2026

Amazon Research Awards recipients announced original ↗

27 May 2026

Diverse reasoning traces teach LLMs to make better decisions original ↗

26 May 2026

Making LLMs faster without sacrificing accuracy original ↗

15 May 2026

Promptimus: Improving already good LLM prompts with zero manual engineering original ↗

14 May 2026

Navigating uncertainty in Amazon's middle-mile network original ↗

6 May 2026

How mechanism design theory helps optimize Amazon-vendor collaboration original ↗

5 May 2026