Aws Re:invent 2025 - Enterprise Ai: Now, Next, Future (ant202)

🦄 Making great presentations more accessible. This project aims to enhances multilingual accessibility and discoverability while maintaining the integrity of original content. Detailed transcriptions and keyframes preserve the nuances and technical insights that make each session compelling.

📖 AWS re:Invent 2025 - Enterprise AI: Now, Next, Future (ANT202)

In this video, Chris Hillman from Teradata and a representative from NVIDIA present their collaborative solution for an automotive manufacturer facing long R&D cycles. They demonstrate how vector embeddings using NVIDIA's Nemotron services combined with Teradata's native vector store enable semantic search of specification documents while processing terabytes of real-time sensor data from test tracks. The system uses AWS Bedrock for private generative AI, allowing junior engineers to access insights previously available only to senior engineers. The Nemotron models achieve 50% more accurate results than open source options, parsing 30 pages per second per GPU. They discuss future development toward AI agents using frameworks like Crew and LangChain, emphasizing the importance of guardrails models and MCP protocols. The solution addresses tool overload through task-specific MCP servers and supports BYOM (Bring Your Own Model) via ONNX format for hybrid cloud and on-premises environments.

; This article is entirely auto-generated while preserving the original presentation content as much as possible. Please note that there may be typos or inaccuracies.

Thanks for taking the time. I need to apologize for being fluffy—it's cold in here and I was wearing an AWS jacket, so I'm all covered in fluff. That's not my normal appearance, but anyway, I'm Chris Hillman, Global AI Lead for Teradata. I'm joined by someone from a small company you might have heard of called NVIDIA. I lead strategic product partnerships and I'm really excited to work with the Teradata team. We're very excited to show you what we built together. Chris, why don't you kick it off and we'll go from there.

Sure. I have a big team and we work in the Americas, EMEA, and APJ. We've been doing great work recently with some of the biggest companies because of the nature of our platform and the huge data sets involved. I've picked out one case to talk to you about today. It's actually related to automotive manufacturing, but hopefully you'll see this is a pattern that's really applicable across all industries, and you'll be able to see why that

Source: Dev.to