engblogs

Mobius Labs' Aana multimodal AI models enhance Dropbox Dash by enabling efficient, context-aware understanding and search of rich media content across text, images, audio, and video at scale.

OpenAI’s Japan Economic Blueprint outlines a strategic framework for leveraging AI to drive inclusive innovation, infrastructure development, and lifelong learning, positioning Japan as a global leader in the AI-driven economy.

OpenAI partners with the UK Government to enhance sovereign AI capabilities, introduce UK data residency, and expand ChatGPT use across public sector to boost productivity and secure AI adoption.

Five individuals with MIT affiliations, including faculty Dina Katabi and Facundo Batista, were elected to the National Academy of Medicine in 2025 for their pioneering contributions to health, digital technology, immunology, and vaccine development.

Lambda leverages NVIDIA GB300 NVL72 systems with liquid-cooled, high-density racks and advanced architectures to build gigawatt-scale AI factories capable of efficiently training and inferring trillion-parameter models and large-scale reasoning workloads.

Corrosion is a distributed state synchronization system employing a gossip protocol and CRDT-powered SQLite to provide a scalable, globally synchronized routing database without relying on distributed consensus.

Metagenomi leverages AWS Inferentia and the Progen2 protein language model to cost-effectively generate millions of novel enzyme variants at scale, reducing computational expenses by up to 56% while maintaining model accuracy.

Leverage Amazon Nova Premier and an agentic multi-agent workflow with Amazon Bedrock Converse API to automate and optimize the migration of legacy C codebases to modern Java/Spring applications while ensuring code quality, security, and integration.

A detailed guide to building a serverless, proactive cost management system for Amazon Bedrock using AWS Step Functions, CloudWatch, and DynamoDB to enforce token usage limits and optimize generative AI expenses.

Explore advanced cost monitoring, custom tagging, and reporting strategies to proactively manage and optimize Amazon Bedrock generative AI expenses using AWS tools like CloudWatch, Lambda, and Cost Explorer.

Leverage Amazon Bedrock and AWS services to build secure, scalable generative AI solutions that streamline and empower product teams' creative workflows across text, image, and audio content.

Half-Quadratic Quantization (HQQ) offers a fast, calibration-free approach to quantizing large machine learning models, achieving competitive accuracy and significant compression speed-ups compared to existing methods.