The awesome power of artificial intelligence offers the potential to solve problems across the globe in mere seconds, from solving complex mathematical equations to diagnosing medical cases.

With great power comes great infrastructure, however, and companies globally are scratching their heads trying to figure out how to utilize the full power of AI, especially without breaking the bank. With Katanemo, tech leaders can develop API-first applications without infrastructure tax while offering privacy and security features for its users, all using intelligent infrastructure, according to Salman Paracha (pictured), founder and chief executive officer of Katanemo Labs Inc.

“We’re building this notion of intelligent infrastructure for developers, starting with a very specific service in this area to solve this perennial challenge around security and observability and around model choice,” he said. “That’s what’s going on in Katanemo at the moment as we play offense with small large language models rather than the ultra-large LLMs.”

Paracha spoke with theCUBE Research analysts Dave Vellante and John Furrier at the “Supercloud 6: AI Innovators” event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed data hygiene in AI, how the company approaches building its infrastructure and how companies are running AI at a massive scale at low costs.

Building the intelligent infrastructure that powers the future

The future of applications is promising with generative AI supporting development, with applications soon being able to accomplish critical tasks, such as updating ad campaigns and insurance requests, more meaningfully. To accomplish this, businesses need to practice good data hygiene, ensuring the data being used is accurate and useful, Paracha pointed out.

“Data hygiene is important, making sure that you have the right cure set that you’re going to use and train models or use in context learning and then be able to put this stack together that accelerates your development,” Paracha said. “That’s what is missing in the market in general, but I think people are catching up to this idea that we need a stack representation, which gives the developers new powers that focus.”

Predicting what will change is next to impossible, so companies need to take a different approach to building intelligent infrastructure: focusing on perennial needs instead. As an example, the rate of large language models being introduced into the market is hard to keep up with, requiring infrastructure that allows endless LLM testing, Paracha explained.

“In the early days, we’ve taken a very perspective view of the type of infrastructure we’re going to build, which is this proxy service, this open-source project that’s going to be announced very soon, and that uses small language models to harness the full potential of large language models,” Paracha said. “I think that will be the constant bet of this venture, which is focusing on what is going to be perennial, not what’s going to change tomorrow.”

Katanemo is as operating at the intersection of science and hypserscaling, and Paracha believes that’s how the next generation of infrastructure is going to be built. With both working efficiently hand-in-hand, the full capabilities of AI can continue to be explored.

“We’re building this first layer of this intelligent prompt-native proxy, an open-source project that we will make available soon to developers so that we can get started on the journey of inventing the new intelligent infrastructure,” Paracha said.

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE Research’s coverage of the “Supercloud 6: AI Innovators” event:

Photo: SiliconANGLE