1p9LQNNXgs2B5cBeLUijT8g

Inside the AIPCon 8 Demos Redefining the Future of Enterprise AI

Editor’s Note: AIPCon 8, Palantir’s most recent customer conference, featured breakthrough customer implementations that demonstrate what’s possible with enterprise AI today. In part one of this two-part series, we share highlights from the afternoon’s standout demo sessions. Those who attended AIPCon 8 all walked away with a shared experience — seeing firsthand the power of transformative AI …

ML 19489a 2

Enhance agentic workflows with enterprise search using Kore.ai and Amazon Q Business

This post was written with Meghana Chintalapudi and Surabhi Sankhla of Kore.ai. As organizations struggle with exponentially growing volumes of data distributed across multiple repositories and applications, employees lose significant time—approximately 30% according to the International Data Corporation (IDC)—searching for information that could be spent on higher-value work. The complexity of modern enterprise data networks …

maxresdefault

Building on the bananas momentum of generative media models on Google Cloud

It’s been exciting to see the capabilities of Nano Banana, our latest image editing model available in Gemini 2.5 Flash Image, go viral. And with transformative workflows like these, it is easy to see why: genmedia bundle carousel 1 Iterative refinement with Gemini 2.5 Flash Image genmedia bundle carousel 2 Context aware conversational editing with …

hapag eta image 1

How Hapag-Lloyd improved schedule reliability with ML-powered vessel schedule predictions using Amazon SageMaker

This post is cowritten with Thomas Voss and Bernhard Hersberger from Hapag-Lloyd. Hapag-Lloyd is one of the world’s leading shipping companies with more than 308 modern vessels, 11.9 million TEUs (twenty-foot equivalent units) transported per year, and 16,700 motivated employees in more than 400 offices in 139 countries. They connect continents, businesses, and people through …

image 1 croppedmax 1000x1000 1

Gemini CLI extension for PostgreSQL in action: Build a fuzzy search feature in minutes

Adding features to an app can be hard. One minute you’re writing code, the next you’re switching to the PostgreSQL database client to run a query, and then it’s over to the console to check on your instances. For example, let’s say you wanted to add search capabilities. This can mean adding the right extensions …

Compute-Optimal Quantization-Aware Training

Quantization-aware training (QAT) is a leading technique for improving the accuracy of quantized neural networks. Previ- ous work has shown that decomposing training into a full-precision (FP) phase followed by a QAT phase yields superior accuracy compared to QAT alone. However, the optimal allocation of compute between the FP and QAT phases remains unclear. We …

ML 19181 Arch

Modernize fraud prevention: GraphStorm v0.5 for real-time inference

Fraud continues to cause significant financial damage globally, with U.S. consumers alone losing $12.5 billion in 2024—a 25% increase from the previous year according to the Federal Trade Commission. This surge stems not from more frequent attacks, but from fraudsters’ increasing sophistication. As fraudulent activities become more complex and interconnected, conventional machine learning approaches fall short …

2 architecturemax 1000x1000 1

Forecasts and data insights come to BigQuery’s MCP and Agent Development Kit tools

For AI agents to be really useful, they need to be able to securely interact with enterprise data. In July, we introduced a toolset to help AI agents interact with and analyze business data in BigQuery through natural language, and with just a few lines of code. Today, we’re taking the next step, with “Ask …

Enabling Differentially Private Federated Learning for Speech Recognition: Benchmarks, Adaptive Optimizers, and Gradient Clipping

While federated learning (FL) and differential privacy (DP) have been extensively studied, their application to automatic speech recognition (ASR) remains largely unexplored due to the challenges in training large transformer models. Specifically, large models further exacerbate issues in FL as they are particularly susceptible to gradient heterogeneity across layers, unlike the relatively uniform gradient behavior …

1q871tK1C7Y8VSAXVOmu4Ig

100X Faster: How We Supercharged Netflix Maestro’s Workflow Engine

By Jun He, Yingyi Zhang, Ely Spears TL;DR We recently upgraded the Maestro engine to go beyond scalability and improved its performance by 100X! The overall overhead is reduced from seconds to milliseconds. We have updated the Maestro open source project with this improvement! Please visit the Maestro GitHub repository to get started. If you find …