Training AI agents to click, type, and navigate real websites requires more than just a great model. Amazon's AGI Lab shares a practical recipe for scaling reinforcement learning into a reliable engine for computer-use agents. https://amzn.to/4vA60CD
Amazon Science
Research Services
Seattle, Washington 388,147 followers
The latest news and research from Amazon’s science community. #AmazonScience
About us
Amazon Science gives you insight into the company’s approach to customer-obsessed scientific innovation. Amazon fundamentally believes that scientific innovation is essential to being the most customer-centric company in the world. It’s the company’s ability to have an impact at scale that allows us to attract some of the brightest minds in artificial intelligence and related fields. Our scientists continue to publish, teach, and engage with the academic community, in addition to utilizing our working backwards method to enrich the way we live and work. Follow us on LinkedIn and visit our website to get a deep dive on innovation at Amazon, and explore the many ways you can engage with our scientific community. #AmazonScience
- Website
-
https://www.amazon.science
External link for Amazon Science
- Industry
- Research Services
- Company size
- 10,001+ employees
- Headquarters
- Seattle, Washington
- Founded
- 2020
- Specialties
- Artificial Intelligence, Machine Learning, Computer Vision, Cloud, Economics, Sustainability, AI, ML, Conversational AI, Natural Language Processing, NLP, Robotics, Security, Privacy, Information, Knowledge Management, Operations, Scientific Research, Search, Amazon, and Alexa
Updates
-
Bringing a drug to market takes 10 to 15 years and costs over $2 billion on average. Using Amazon Nova Forge, Nimbus Therapeutics built its own customized frontier model, Novus, a scientific assistant trained in chemistry that will reason through molecular design. Through supervised and reinforcement fine tuning, one LLM now matches the performance of multiple specialized graph neural networks, unlocking the future potential of AI in drug discovery. https://amzn.to/4emDPki
-
-
Therapeutic antibody discovery remains slow and resource-intensive, with traditional methods providing limited control over epitope selection. We present a workflow for de novo nanobody design applied to a novel Desmoplastic Small Round Cell Tumor target encompassing four stages: (1) epitope identification guided by our hotspot recommendation agent using physical chemistry-based structure and sequence analysis tools with two curated databases (IEDB, PFAM), (2) de novo nanobody generation using three independent methods (RFantibody, IgGM, mBER) across multiple predicted antigen structures and nanobody frameworks, (3) multi-metric scoring including structural metrics from folding models, and in silico binding affinity from our sequencebased predictor, (4) high-throughput yeast surface display (YSD) screening followed by surface plasmon resonance (SPR) characterization of the specific binders. We generated 288,000 nanobody designs spanning eight target epitope regions and three variable domains of heavy chain-only antibody (VHH) frameworks. Multi-objective Pareto filtering with our candidate selection agent yielded 100,000 candidates for YSD screening with fluorescence-activated cell sorting (FACS). Of 116 enriched candidates advanced to SPR characterization, 46/116 (39.7%) produced reliable kinetic fits with Rmax ≥ 30 RU, yielding KD values from 0.66 nM to 305 nM (median 31.7 nM). These results show that an agent-guided computational workflow can design nanomolar to sub-nanomolar nanobody binders against a novel target without experimental structure or prior antibody information.
-
Amazon built RuleForge to tackle the challenge of 48,000 new vulnerabilities published in 2025. The agentic AI system generates production-ready detection rules 336% faster than manual methods: https://amzn.to/4dzAnCJ
-
-
Amazon Science reposted this
Millions of Uber trips start with a tap. Behind each one: AWS infrastructure processing matches in milliseconds. 🚙 Now, Uber is expanding on AWS—using Graviton4 for real-time Trip Serving Zones and piloting Trainium3 to train AI models for driver matching, arrival predictions, and personalized recommendations that will scale globally. "Uber operates at a scale where milliseconds matter," said Kamran Zargahi, VP of Engineering at Uber. The result: faster matches, smarter experiences, and reliable service for hundreds of millions of daily users worldwide. Check out the details: https://go.aws/41fAmwr
-
-
Intelligence is about time, not scale. AWS researchers prove that minimizing inference time maximizes the algorithmic mutual information between training data and future tasks. One formula explains why: https://amzn.to/411Lltc
-
-
Amazon Research Awards Spring 2026 call for proposals is now open across seven research areas, including Agentic AI and Robotics. Successful applicants receive unrestricted funds, AWS promotional credits, and training resources. Deadline for submissions is May 6: https://amzn.to/3PY9XAw
-
