DBRX at Data + AI Summit: Best Practices, Use Cases, and Behind-the-scenes

Businesses are making remarkable progress on their data and AI journeys. They’re advancing from a few pilot projects confined to use cases likely deemed “non-mission-critical,” to deploying applications in real-world operations across thousands of users that are enhancing employee productivity and improving efficiency. 

However, building and scaling these next-generation systems is a complex, resource-intensive combination of engineering challenges up and down the stack. And as more users look to build their own GenAI applications, they’re encountering many of the same issues. 

At Databricks, we have learned quite a bit about those challenges firsthand. We used our Data + AI Platform to train our own foundation model, DBRX. The lessons we learned in that process have been baked into our infrastructure, making it resilient and scalable, while delivering world-class performance.

So, while DBRX is a fantastic foundation model for building high-quality, efficient AI applications, that’s just one part of the story. The other critical part is in the infrastructure, and getting all the details right. When you use Databricks to train, fine-tune, and deploy AI models for your own needs, you use the same platform we do. In that way, the most important benefit of our experience is that our customers get the incredible resilience, scalability, and proven performance of the Data + AI Platform.

At the upcoming Data + AI Summit, attendees will have the opportunity to learn how to build AI-based applications with Databricks, including how they can take advantage of DBRX and other foundation models.

Below are a few of the sessions focused on building and using foundation models. And be sure to check out the keynote on Wednesday, June 12, to hear how Databricks is making GenAI a reality for enterprises.

Building a production-scale, private, OSS RAG pipeline with DBRX, Apache Spark™ and LanceDB

Speakers: Chang She (Co-Founder, LanceDB), Jasmine Wang (LanceDB)

One of the biggest hurdles to enterprise AI adoption is data security. This session will walk through how combining Spark, DBRX, and LanceDB can help companies build Generative AI projects without putting sensitive information and IP at-risk. 

Session info here.

In the trenches with DBRX: Building a state-of-the-art open source model

Speakers: Jonathan Frankle (Databricks), Abhinav Venigalla (Databricks)

Want to learn the behind-the-scenes story on how Databricks built DBRX? Join this session to hear from the team behind it about the tools and methods used, as well as the lessons they learned, to help you build your own state-of-the-art LLM. 

Session info here.

Introduction to DBRX and other Databricks foundation models

Speakers: Margaret Qian (Databricks), Hagay Lupesko (Databricks) 

This session will provide an introduction to the foundational models available on Databricks, including DBRX. Participants will leave with a better understanding of the capabilities and applications of these models, and how to apply them to drive business value.

Session info here.

7 DBRX hacks no one told you about

Speakers: Indivar Nayyar (LTIMindtree)

With the recent launch of DBRX, Databricks has significantly elevated their GenAI offerings. This session will cover real-world examples of how companies are already using DBRX in innovative ways. This event is sponsored by LTIMindtree. 

Session info here.

 

Register for 2024 Data + AI Summit here! Join us in person in San Francisco, or participate virtually for free.

---

Frequently asked questions

What is DBRX and how did Databricks build it?

DBRX is Databricks' own foundation model, trained using the company's Data + AI Platform to prove out the infrastructure before making it available to customers. The engineering lessons learned while training DBRX were fed back into the platform, making it more resilient and scalable while still delivering strong performance. Readers can find the original technical deep dive in the DBRX launch post.

Do customers get the same infrastructure Databricks used to train DBRX?

Yes, when you use Databricks to train, fine-tune, and deploy AI models, you use the same Data + AI Platform that Databricks used to build DBRX. This means customers benefit directly from the resilience, scalability, and proven performance the Databricks team validated while building its own foundation model. That shared infrastructure is presented as the core value proposition for anyone building GenAI applications on Databricks.

Which Data + AI Summit sessions focus specifically on DBRX?

Four sessions highlighted in the article center on DBRX from different angles: building a private RAG pipeline with DBRX, Apache Spark, and LanceDB; an insider look at how the Databricks team built DBRX in "In the Trenches with DBRX"; a broader overview in "Introduction to DBRX and other Databricks Foundation Models"; and a sponsored LTIMindtree talk on real-world DBRX use cases called "7 DBRX Hacks No One Told You About!". Each session approaches DBRX from a different practical or technical perspective, from security to hands-on tips.

What will the RAG pipeline session with LanceDB cover?

The session, "Building a Production Scale, Totally Private, OSS Rag Pipeline with DBRX, Apache Spark, and LanceDB," walks through how combining Spark, DBRX, and LanceDB helps companies build generative AI projects without exposing sensitive information or IP. It is presented by LanceDB co-founder Chang She and Jasmine Wang, and it directly addresses data security as one of the biggest hurdles to enterprise AI adoption.

How and when can I register for Data + AI Summit 2024?

You can register for Data + AI Summit through the official registration page, with the option to attend in person in San Francisco or participate virtually for free. The article also points attendees to the keynote on Wednesday, June 12, to hear how Databricks is making GenAI a reality for enterprises.