"The AI arms race is a zero sum game... 10 years from now, the Fortune 500 will look very different because some companies did not get into collecting the data and training their machine learning models early enough."
Who are we?
KatenAI is an MLOps company focused on solving Kubernetes(K8s) optimization problems, starting with capacity allocation and cost attribution as building blocks to a much bigger end game. Today, this is highly manual, ineffective, and expensive for customers using cloud resources.
We are building a recommendation engine based on machine learning of workload behavior in Kubernetes clusters. We optimize workload distribution, cluster capacity resources, and your costs across cloud providers.
Our Mission
We want to tilt the power towards the consumers of cloud resources so, every workload can be optimized using the 3 levers of cost, performance, and capacity allocation at the right time, at the right price, in the right cloud.
Before Kubernetes, these optimizations were not easily solvable, now they are, using open source technologies.
Kubernetes is mature enough today to solve problems above the stack.
“A lot of times people are building products on top of Kubernetes and they focus too much on Kubernetes.” - @kelseyhightower
Are you happy with what you are paying your cloud provider? Is it justified?
Today, consumers of cloud resources have no negotiating power with cloud providers. It’s very easy to get started but then you are stuck. Every month you look at your bill and get sticker shock. The pain is acutely felt when your bill goes above $3K/month.
Unlike billing optimizations, it’s not that the cost is good or bad. It may be reasonable what you are paying, but how do you know if what you are paying is what you should be paying? There’s no way to know and no way to justify it. Spot instances can save you money if your workloads are relatively static and your apps are monoliths.
If your workloads are built as microservices in containers, deployed frequently, then the dynamic and real-time nature makes this a challenge for humans to solve. The problem gets exponentially worse the more workloads, clusters, and regions where your applications are distributed.
That’s where we can help.
Coming soon… a story of a cloud provider explaining how to save on $250K monthly spend by using spot instances. Care to guess what the outcome was?
It’s not just the cost, it’s the cost/performance ratio that matters.
@451Research - cost/performance optimization is the #1 need