As data continues to be the backbone of enterprises, data visualization has become critical for the easier identification of outliers, trends and patterns in huge data sets.

Nomic Inc. makes this a reality through a product called Nomic Atlas that organizes unstructured data using a map that is powered by artificial intelligence. In the process, data exploration and filtering is enhanced, according to Brandon Duderstadt (pictured, right), founder and chief executive officer of Nomic.

“We call that the map, and that’s part of our product, Nomic Atlas,” Duderstadt said. “In Nomic Atlas, every piece of your data set is a point on a scatter plot, and two points are close if the content of that data is similar. We can apply a lot of different visualization techniques on top of it.”

Duderstadt and Suraj Patel (left), head of MongoDB Ventures at MongoDB Inc., spoke with theCUBE Research’s chief analyst Dave Vellante at the MongoDB.local NYC event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed the importance of data visualization in this day and age, as well as how Nomic helps MongoDB with this objective. (* Disclosure below.)

MongoDB’s Suraj Patel and Nomic CEO Brandon Duderstadt talk with theCUBE about the importance of data visualization.

Data visualization needs cement MongoDB/Nomic partnership

Based on MongoDB’s mission of empowering innovators to use software and data to disrupt industries, the company inked a partnership with Nomic to enhance software developers’ productivity. This is because the data visualization deployed helped address burning issues, such as fine tuning a model and putting a retrieval-augmented generation system into production, Patel pointed out.

”It was actually a little bit of a different thesis when we invested in Nomic,” he noted. ”Certainly, the rubber hit the road with ChatGPT’s release in 2022. When we met Nomic a few months later, we were seeing this phenomena of not necessarily subject matter experts, but software developers that really wanted to understand what’s going into models, how are models being trained. And Nomic had this really powerful visualization.”

Bringing AI into everyday applications continues to be top of mind for enterprises. As a result, Nomic makes this a reality through its open-source software platform, called GPT4ALL, that enables the deployment and training of customized and powerful large language models on everyday hardware, according to Duderstadt.

”One of the things that we think a lot about at Nomic, this goes into our third offering, GPT4ALL, is the unit economics of these models,” he said. ”GPT4ALL comes into play as a platform that allows us to take these specialized models and compress them and run them either in the cloud on hardware that is more efficient or on the edge on a user device.”

As generative AI models continue to take center stage, as evidenced by ChatGPT, RAG has become a fundamental concept. As a result, retrieving the right data is an important part of RAG, and this is why Nomic Embed comes in handy, according to Duderstadt.

”One of the things that Nomic has done is we’ve released this model, Nomic Embed. ‘It’s the first truly open and data model that outperforms OpenAI’s embedding models on both short and long context tasks,” he said. “We’ve seen massive adoption of this thing. We just passed a million downloads on Hugging Face, and we’re really excited about enabling the entire broader developer community to do RAG with their fully open-source stack, powered by Nomic.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE Research’s coverage of the MongoDB.local NYC event:

(* Disclosure: TheCUBE is a paid media partner for the MongoDB.local NYC event. Neither MongoDB Inc., the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE