NeuBird founders Goutham Rao and Vinod Jayaraman came from PortWorx, a cloud-native storage solution that they eventually sold to PureStorage in 2019 for $370 million. This was their third successful sending off.
Last year, when I was looking for my next startup challenge, I saw an opportunity to combine my cloud-native knowledge, especially in IT operations, with the fast-growing field of generative AI.
Today, Neubird announced a $22 million investment from Mayfield to take the idea to market. That's a lot of money for an early-stage startup, but the company is likely banking on the founders' previous experience to build another successful company.
CEO Rao says the cloud native community has done a good job in solving many difficult problems, but in the process has created an increasing level of complexity.
“We've done a great job as a community over the past decade building cloud-native architectures with service-oriented design. This has added a lot of layers, and that's a good thing. is a good way to design software, but it came at the cost of more telemetry – too many layers in the stack,” Rao told TechCrunch.
They concluded that this level of data makes it impossible for human engineers to discover, diagnose, and solve problems at scale within large organizations. At the same time, large-scale language models were beginning to mature, so the founders decided to use them to tackle the problem.
“We are uniquely leveraging large-scale language models to analyze thousands of metrics, alerts, logs, traces, and application configuration information in seconds, allowing you to diagnose the health of your environment. , detect if there is a problem and come up with a solution,” he said.
The company is essentially building a trusted digital assistant for engineering teams. “So this is a digital co-worker who works alongside his SRE and ITOps engineers and monitors all alerts and logs looking for issues,” he said. The goal is to reduce the time it takes to respond and resolve incidents from hours to minutes, and we believe that leveraging generative AI to problems can help businesses achieve that goal.
The founders understand the limitations of large-scale language models, and by training the model using limited datasets and setting up other systems to help provide more accurate responses. We are trying to reduce hallucinations and incorrect responses.
“We are using this in a very controlled way for very specific use cases in environments that we know about, so we cross-check the results we get from the AI again through our vector database. , and see if it makes sense. If you're not happy with it, we won't recommend it to you.”
Customers can connect directly to various cloud systems by entering their credentials, and NeuBird can use that access to match other available information to find solutions, without moving any data. , reducing the overall difficulty associated with obtaining company-specific data. For models to work with.
NeuBird uses various models, including Llama 2, to analyze logs and metrics. They use Mistral for other types of analysis. The company actually converts all natural language interactions into SQL queries, essentially turning unstructured data into structured data. They believe this improves accuracy.
The early-stage startup is currently working with design and alpha partners to refine the idea in order to bring the product to market later this year. Rao says they raised so much money so quickly because they wanted to create a room to work on the problem without worrying about finding more money right away.