From telemetry data to geospatial queries, today’s fleets demand intelligent systems to keep operations running smoothly. MongoDB makes it possible. This demo explores how an AI agent, powered by MongoDB Atlas and Atlas Vector Search, simplifies complex decision-making for fleet managers — delivering fast, actionable insights when they matter most. 👇 https://lnkd.in/e9NV8d58

From telemetry data to geospatial queries, today’s fleets demand intelligent systems to keep operations running smoothly. MongoDB makes it possible. This demo explores how an AI agent, powered by MongoDB Atlas and Atlas Vector Search, simplifies complex decision-making for fleet managers — delivering fast, actionable insights when they matter most. 👇 https://lnkd.in/e9JDFTFm

If you're looking to hear about the intersection of Big Data and AI on the cutting edge, The Trade Desk London is hosting a set of talks with our ClickHouse vendor Altinity, Inc. on the 23rd of October. (https://luma.com/dvgbplae) CEO Robert Hodges is speaking and bringing together a great roster from Databricks, Amazon Web Services (AWS) and Confluent covering: - Project Antalya: Altinity’s initiative to adapt ClickHouse OSS for decoupled storage and compute, using Iceberg as the open table format. Designed to be performant, cost-efficient, and open source—at AI data scale. - Confluent: Exploring how to interact with databases using natural language, leveraging a wide range of components. Learn about key concepts like MCP, LLMs, and NL2SQL that make this possible. - AWS: S3 Table Buckets—an evolution of S3 that natively supports Iceberg tables, offering a rich set of managed features and optimizing S3 for Iceberg query performance. Accessible to a variety of compute engines including ClickHouse as mentioned. - Databricks: Topic to be revealed... We're excited for The Trade Desk to host this event and can't wait to see the architectural seeds it plants for the future. If this sounds interesting, please join us—and feel free to share ❤️

From telemetry data to geospatial queries, today’s fleets demand intelligent systems to keep operations running smoothly. MongoDB makes it possible. This demo explores how an AI agent, powered by MongoDB Atlas and Atlas Vector Search, simplifies complex decision-making for fleet managers — delivering fast, actionable insights when they matter most. 👇 https://lnkd.in/dtG7BaaC

Cutting Edge Growth Hack for GenAI 🚀 How do you keep your Retrieval-Augmented Generation (RAG) systems humming with fresh data — without breaking the bank or waiting forever on ingestion? 👉 Enter Delta Lake Change Data Feed (CDF) on Databricks. Here’s the deal: • Instead of rebuilding your entire vector store every time new data arrives, CDF lets you *incrementally* update only what’s changed. • The result? Dramatically slashed ingestion costs and latency. • Your RAG pipeline keeps data *fresh* and *accurate* in production — essential for real-time AI applications. This isn’t just theory — organizations running large-scale RAG deployments on Databricks report *significant savings* in cloud bills and speed-to-insight. If your GenAI stack still does batch-heavy vector store refreshes, maybe it’s time for a rethink. What’s your approach to balancing freshness, cost, and speed in vector search? Curious to hear real-world tactics! #GenAI #Databricks #DeltaLake #RAG #VectorSearch #DataEngineering #MachineLearning #AIInnovation

How I Built an Autonomous AI Customer Retention Agent with AWS Bedrock AgentCore Built for the AWS AI Agent Global Hackathon After building a serverless data analytics pipeline for customer churn, I had clean, query-ready customer data sitting in Amazon Athena. The next logical step was to make that data actionable — not just for analysts, but for customers themselves. That's where the Customer Retention Agent comes in. This is a fully autonomous AI agent built on AWS Bedrock AgentCore that identifies at-risk customers and proactively offers them personalized retention deals through natural conversation. I built this as part of the AWS AI Agent Global Hackathon, and it's a natural continuation of my previous project. Before diving into the build, I spent time going through the Amazon Bedrock AgentCore Samples repository. The tutorials there were incredibly helpful for getting up to speed with AgentCore concepts — from Runtime and Gateway to Memory and Identity. If you're new to AgentCore, I highly recommend starting there. The goal was simple: What if customers cou https://lnkd.in/gRwb89gH

Fine-tuned (PEFT/LoRA) models can greatly outperform out-of-the-box LLMs, but serving them robustly can be challenging due to both numerical issues and performance issues. Our engineering and research teams at Databricks have done amazing work to get up to 2x higher throughput *and* 2-3% better accuracy than open source serving engines for LoRA. Here are some of the techniques we developed:

When we think about AI competition, it is easy to assume the winners will be decided by who has the biggest model. But as MongoDB's Scott Sanchez and Pete Johnson explain in The New Stack, the real battle is being fought in go-to-market strategies. From pricing models to safety guardrails, execution choices are defining the industry just as much as new technology. Check out the article for their analysis and three key insights. 👇 https://lnkd.in/gRASMzXA

Turning millions of Amazon reviews into decisions with LLMs and Airflow. This episode of “The Data Flowcast: Mastering Apache Airflow® for Data Engineering and AI” features Naseem Shah at XENA Intelligence . We discuss moving from cron to Airflow, balancing batch and concurrency, and structuring text into strategy. Click the link in the comments for the episode. #AI #Automation #Airflow #MachineLearning

When we think about AI competition, it is easy to assume the winners will be decided by who has the biggest model. But as MongoDB's Scott Sanchez and Pete Johnson explain in The New Stack, the real battle is being fought in go-to-market strategies. From pricing models to safety guardrails, execution choices are defining the industry just as much as new technology. Check out the article for their analysis and three key insights. 👇 https://lnkd.in/dFyfYarX