engblogs

AI-driven optimization of the power grid enables real-time balancing of supply and demand, better integration of renewables, and smarter planning to improve efficiency, resilience, and decarbonization.

Scaling JAX-powered LLM training from a single GPU to multi-node NVIDIA Blackwell clusters, with controlled experiments, learning-rate scaling, and practical patterns for ML engineers.

A concise technical guide on how to scale agentic systems within enterprise environments, drawing on Netomi’s practical lessons.

A technical synthesis of Arctic diagnostics and AI-powered subseasonal forecasting that extends winter-weather predictions beyond ENSO, highlighting Siberian snow cover, polar-vortex dynamics, and Barents-Kara Sea conditions.

A technical overview of applying OpenAI in healthcare, focusing on model integration, data governance, and safe, AI-assisted clinical workflows.

A PoC for AI-assisted, scalable medical content verification at Flo Health using MACROS on Amazon Bedrock to streamline content review and revision against up-to-date medical guidelines.

FINRA leverages Google Cloud's DORA program to operationalize data-driven DevOps at scale, driving enterprise-wide modernization and faster production updates through a culture of continuous improvement and data-first decision making.

Demonstrates how OLAF, integrated with Locust, load-tests SageMaker inference endpoints to reveal latency and throughput bottlenecks and enable scalable, cost-aware ML deployments with Observe.AI’s testing tool.

Automates PII detection and redaction across text and image content in emails and attachments using Amazon Bedrock Data Automation and Guardrails, with secure data handling, auditing, and a web portal for review at scale.

Lambda appoints Leonard Speiser as Chief Operating Officer to lead day-to-day operations and accelerate deployment and scaling of its AI infrastructure-powered supercomputing platform for enterprise customers.

Unified security analytics for energy and utilities that blends IT and OT telemetry into a data lakehouse using Databricks to enable real-time threat detection and automated compliance.

Metadata-driven templating with DLT-META turns Spark Declarative Pipelines into scalable, governance-ready workflows by centralizing templates and auto-generating pipeline logic.