MLnative

MLnative is a platform for running Machine Learning models in production, delivering 10x improvement in resource utilization and cost efficiency. Our solution provides: GPU sharing, autoscaling, customizable priority queues and user-friendly interface for ML models deployment and management. Our platform can be deployed either on the cloud resources or on-premise infrastructure. It is installed in your environment, so you keep everything under control.

Key Features

• GPU sharing

• Autoscaling

• Customizable priority queues

• Easy deployments

• Web app and REST API

FAQ

How does it work?

MLnative provides the customer with a dedicated platform, available via a set of intuitive UI and programming APIs for managing models in production. The platform leverages a range of open-source technologies, as well as a handful of proprietary tweaks to maximize GPU utilization and scalability.

Does my data leave the company network?

Our clusters are fully isolated - there is no communicattion with external services. None of your data is ever leaving your servers.

Who manages the infrastructure?

MLnative manages the infrastructure on the customer's resources, whether that's on any of the supported public clouds, or on-premise.

What does the support look like?

We provide full docs on how to work with the platform, end-to-end example integrations for reference (e.g. a Text-to-speech application built upon MLnative), and a dedicated per-customer support slack channel. We support our customers very actively during the initial onboarding period so that the onboarding process is as smooth as possible.

Do you support air-gapped environments?

Yes, a complete hands-off approach in case of the most demanding security concerns is available. We provide our customers with installation packages, guidance, and instructions on how to run MLnative effectively.