engblogs

Learn how to build your first AI-powered travel chatbot using Google's Gemini model and Vertex AI SDK, including system instructions and real-time data integration with function calling.

Google Cloud introduces Axion C4A metal, a bare metal Arm-based instance designed for specialized workloads requiring direct hardware access and architectural parity across automotive, Android development, and other critical use cases.

Google Cloud's new Axion-based N4A VMs deliver up to 2x better price-performance and enhanced flexibility for general-purpose and AI workloads, featuring custom machine types, advanced Hyperdisk storage, and integration across Compute Engine, GKE, and more.

Google Cloud launches Ironwood TPUs and new Arm-based Axion VMs, delivering unprecedented performance, efficiency, and scalability for AI inference and general-purpose compute workloads.

An in-depth exploration of Google's Ironwood AI stack showcasing the co-designed hardware and software architecture, including TPUs, XLA compiler, JAX and PyTorch frameworks, and advanced tools that deliver exceptional performance, scalability, and efficiency for large-scale AI model training and inference.

Exploring the simplicity and power of building LLM agents with minimal code, highlighting context management, tool integration, and open challenges in agent design.

Cloudflare Stream now enables efficient extraction of high-quality M4A audio tracks from videos via a simple API or dashboard, supporting advanced workflows like AI transcription, translation, and content moderation.

tokio-quiche is an open-source asynchronous Rust library integrating quiche with the Tokio runtime to simplify building high-performance QUIC and HTTP/3 applications.

PolyNorm leverages few-shot prompting with Large Language Models to enhance scalable, multilingual text normalization in Text-to-Speech systems, reducing manual rule crafting and improving performance across diverse languages.

Unify and manage multiple specialized MCP servers seamlessly through Amazon Bedrock AgentCore Gateway, enabling centralized authentication, tool discovery, and operational efficiency for scalable AI agent architectures.

Exploring how the Modular Platform leverages Mojo and MLIR to unify AI software development across diverse hardware and software layers, addressing challenges in performance, portability, and developer experience for the future of AI innovation.

HP Industrial Print revolutionized its data platform using Databricks SQL, achieving 40% faster pipeline performance, enhanced governance, seamless data sharing, and new revenue through scalable data products and monetization.