Intelligent application development startup Clarifai Inc. said today it’s expanding its artificial intelligence compute capabilities with the launch of a new compute orchestration platform that spans every public cloud and on-premises server.

The new capabilities are all about helping organizations boost the performance of their AI applications and optimize spending to get more bang for their buck.

Clarifai is the creator of a full AI stack platform that enables companies to create intelligent applications that leverage vast volumes of unstructured and unlabeled data as a source of knowledge. To that end, it offers multiple application programming interface for tasks such as natural language processing, audio and speech recognition, and more.

With Clarifai’s platform, app developers can train the AI models that power their applications using unstructured data. Its platform features a vast data lake, a Scribe Automated Data Labeling tool for automatically labeling unstructured information, a Spacetime Search tool for indexing that data, and an Edge AI service to layer AI onto data streams generated by cameras and drones.

The new compute orchestration features, available in preview now, are the latest addition to its platform. The company says they help customers maximize their AI investments across every computing environment they use. That provides a place for companies to organize all of their AI compute resources — be it from their own servers, the public cloud or Clarifai’s own software-as-a-service-based compute service, so they can benefit from centralized management and monitoring to optimize costs, performance and governance.

With this, Clarifai says, it’s the only AI platform that can orchestrate AI workloads across any kind of cloud or hardware provider, helping companies to avoid vendor lock-in.

Clarifai founder and Chief Executive Matt Zeiler said the new capabilities should be useful as enterprises continue to experience pains when it comes to deploying AI in production. He said customers want greater flexibility in terms of deployment. To do this,s they need a reference architecture for deploying AI repeatedly, along with a full-stack of AI tools for custom AI workloads, so they can boost performance and reduce costs.

Zeiler said the compute orchestration platform comes with a simple control plane for governing access to AI resources, cost management and performance monitoring. It does this by optimizing resource usage using techniques such as simple dependency management, customizable autoscaling and model packing, in order to reduce AI computing costs by up to 60% on an average workload.

Another advantage is more flexible deployment, with customers easily able to choose the environment where they perform inference for any application. The startup notes that in this regard, it’s one of the few companies to support on-premises inference, and the only company in the world to support inference in air-gapped environments.

Other benefits include Clarifai’s enhanced security. Its AI compute platform makes it possible to deploy the control plane within a virtual private cloud or on-premises Kubernetes clusters, without opening ports to the customer’s environment or creating new infrastructure access management roles for public cloud infrastructures. Finally, the new platform offers native integration with Clarifai’s entire stack of AI development tools, making it easier to customize any AI application.

“We’re opening up capabilities we built over the years to optimize our compute while serving billions of AI requests continuously and serving millions of models simultaneously,” Zeiler said. “Our customers can now have the same world-leading tools to scale their AI workloads on their compute, wherever it may be.”

Image: SiliconANGLE/Stable Diffusion