engblogs

Announcing a global, peer-powered Chrome Enterprise Community that lets IT and security professionals access official resources, share best practices, and securely manage Chrome Enterprise across Windows, ChromeOS, macOS, and Linux.

A deep learning model named PULSE-HF forecasts which heart-failure patients will experience a decline in left ventricular ejection fraction within the next year from ECG data, guiding follow-up and resource allocation.

Explore how Policy in Amazon Bedrock AgentCore enforces deterministic external Cedar policies at the gateway to safely govern AI agents in healthcare, preventing data exposure and unauthorized tool use.

Gain production-ready visibility into Amazon Bedrock inference with new server-side CloudWatch metrics TimeToFirstToken and EstimatedTPMQuotaUsage, enabling model-level latency and quota monitoring, alarms, and capacity planning without API changes.

A technical tour of how Spotify Wrapped Archive combines data pipelines, prompt engineering, model distillation, and automated QA to generate up to five personalized, data-grounded narratives for hundreds of millions of users at global scale.

Chrome arrives on ARM64 Linux in 2026, delivering a secure, Google ecosystem-driven browsing experience with cross-device sign-in, extensions, and AI-powered protections.

LiTo proposes a 3D latent representation that tokenizes surface light fields to jointly model geometry and view-dependent appearance, enabling single-image conditioned generation of 3D objects with realistic lighting, specular highlights, and Fresnel effects via latent flow and differentiable rendering.

Governance as an operating framework—embedding guardrails, traceable architecture, and continuous feedback to enable trusted, scalable AI in large organizations.

A practical guide to designing and evaluating a travel destination recommender system that enhances user exploration and discovery.

Context-aware Sensitive Data Protection (SDP) enables context-driven detection and redaction of sensitive data across AI training and live interactions to improve data security.

A concise guide to building resilient LLM applications on Vertex AI by selecting appropriate consumption models and applying patterns such as global routing, context caching, prompt optimization, and traffic shaping to reduce 429 errors.

Continuous AI enables a cross-team accessibility feedback workflow at GitHub, automating triage and routing feedback into tracked, prioritized issues to drive continuous, human-guided improvements.