PromptOps: How Generative AI Can Help DevOps
What can the Large Language Models and generative AI bring to the world of DevOps? PromptOps has an answer: It can save time and money in the process of automating routine operations.
The company, originally called CtrlStack, is named after its process engine, which is “able to understand human requests and commands and questions, able to read knowledge bases, able to generate the code on the fly, [all in order] to achieve tasks,” said Dev Nag, founder and CEO of PromptOps (yes, that’s his real name), in our latest podcast for The New Stack.
For this podcast, Nag is joined by GK Brar, PromptOps founding engineer, to explain to our audience what generative AI is and how it could help DevOps.
A lot of what PromptOps does is repetitive, Nag admitted, And this is because the daily work of a DevOps professional can involve the same troubleshooting and reporting over and over again. ” If you go to Stack Overflow, you’ll see the same question might have 500,000 views,” Nag noted. So it makes sense to automate these actions wherever possible. PromptOps can do that, without the messy trial-and-error tribulations to work out an automation script.
One thing PromptOps does is intent matching. A DevOps pro may ask PromptAI to “Spin up a cluster.” But what does that mean? A Kubernetes cluster, an RDS cluster, or another cluster entirely? “So we’ll use past history with us to fill in that intent gap: Really, what he’s asking for is Kubernetes clusters and not RDS clusters,” Brar explained.
“Most of our uses of generative AI is actually offline so that it does not impact the experience of the user,” Brar elaborated. Instead, it is used for preparation to automate common sets of actions.
To be fair, there are a lot of application performance monitoring companies and other DevOps start-ups using generative AI to enhance their products. PromptOps is a bit more ambitious, the CEO asserts.
Companies “will tend to staple LLMs onto their products to make them easier to use,” Nag said. “But I think the scope of what’s possible here is just a lot wider. For us, we’re not trying to make one legacy product easier to use, we’re trying to leverage all of DevOps, with this amazing technology.”
Give a listen to this whole interview for other insights.
