Amazon Bedrock AgentCore Introduces Dataset Management for Agent Evaluation
Amazon Bedrock AgentCore now includes dataset management to improve agent evaluation, allowing versioned test scenarios with stable inputs and ground truth assertions.
Archive
Browse all published articles.
ARCHIVE
-
-
AWS Introduces LangSmith for Deep Agent Evaluation
AWS and LangChain collaborate to offer a framework for testing AI agents, using Amazon Nova 2 Lite for evaluation tasks with a 1 million-token context window.
-
AWS Introduces REST API Proxy for SageMaker MLflow Integration
AWS announced a REST API proxy solution to enable secure HTTPS access to Amazon SageMaker MLflow, simplifying integration with existing enterprise systems.
-
Thermal Cameras and AI Help Ships Avoid Gray Whales in San Francisco Bay
Thermal cameras and AI systems now help ships avoid gray whales in San Francisco Bay, reducing collision risks after whales began stopping there during migration in 2026.
-
HuggingFace Explores Genetics-Based AI Agent Prompts
HuggingFace's experiment uses a 120-character genome string to guide AI agent behavior, achieving 68% alignment with expected outputs in testing.
-
Vertu Unveils AI-Foldable Phone for Executives Starting at $6,880
Vertu launched the Alphafold, a foldable phone with AI agent for enterprise workflows, priced at $6,880 for the calfskin version.
-
Social Engineering Good: IEEE's Call for Ethical AI Practices
IEEE urges tech firms to adopt ethical AI practices to combat misuse, citing a 2026 survey showing 68% of engineers express concern over AI's societal impact.
-
MathWorks Demonstrates AI-Based Virtual Sensors for Battery Management Systems
MathWorks' webinar shows how AI virtual sensors can estimate battery state of charge with 95% accuracy in real-time.
-
IEEE Radar System Distinguishes Insect Species
IEEE researchers developed a radar system that can differentiate between insect species with 95% accuracy, offering new tools for tracking pollinator populations.
-
Hugging Face Introduces Trimming Technique for Lightweight Models
Hugging Face's trimming technique reduces model size by up to 25% without performance loss, as demonstrated with GPT2-small.
-
Māori Voice Model Challenges Big Tech Dominance
A Māori-owned text-to-speech model, developed by the University of Waikato, aims to reduce reliance on dominant tech companies' AI systems.
-
Melbourne AI Data Center Boosts Research Innovation
Melbourne’s AI data center, MAVERIC, supports 150 researchers with secure infrastructure for sensitive data analysis, per IEEE.