- Use UbiOps to add data science and AI capabilities to your own website, dashboard or internal data applications.
- Deploy your code with our browser based UI, with the UbiOps CLI or Client Library, or with our API directly
- Support for both Python and R
- Include any external library, artifact, or other code dependency in your deployment folder
All the functionality you need to make your data science application a reality.
UbiOps is an easy-to-use deployment and serving layer for your data science code. Run your Python & R models and scripts live and use them from anywhere at any time.
Deploy your code
Run your code live behind a secure serving endpoint in only a few clicks.
- Set and install OS level dependencies for the deployment container
- Easily create new deployment versions and keep track of revisions
Serving & Operation
UbiOps turns your code into a live, managed service.
- UbiOps takes care of resource allocation, scaling, monitoring and security. Without you having to worry about any of this.
- Make requests to your deployment through our API or by using our WebApp, client libraries and request scheduling tools
- Automatic resource allocation & scaling based on demand
- Support for low-latency synchronous requests as well as asynchronous batch requests
- Keep track of requests and their status
- Gain insight in everything that’s going on by using our log viewer.
- Data input & output options include structured data, files and plain text strings
- Get metrics on usage and performance
- Data schema enforcement and type checking
Create an account & Deploy your code today
Get access to all UbiOps features!
Connect multiple deployments in a pipeline and run them in sequence.
- Create deployments for data transformation, prediction, or pre/post-processing and connect them in any way you need.
- Option to bypass deployments and merge output from multiple deployments to one
- Connect to databases, warehouses and other storage systems using our ready-to-go connector templates
- Pipelines get a single unique serving endpoint
- Scheduling functionality for single deployments and pipelines to run them on set dates & times.
Integration with other tools & services
We support integrations with other parts of your stack through code examples and templates
- Make use of our Client Library and CLI to script the deployment process or build integrations
- See our cookbook for examples and templates for connecting with data storage, training systems and much more.
- Use our templates for Github Actions or Gitlab Pipelines to match your CI/CD process
- Templates and examples for connecting to common data sources and services
- Deploy from training environments like Azure ML services and MLFlow
Security & Governance
Customizable user management and permission controls
UbiOps handles your data with care and only processes the data for the duration of your requests
Fail-safe and self-healing deployments with Kubernetes and KNative
Audit logs as a record for changes to deployments and pipelines
Secret and credential management. No need for hardcoding passwords in your code
SSO authentication for Google and Microsoft Active Directory
We protect your data and code with built-in state-of-the-art security
”I’m a big fan of UbiOps because you can deploy, maintain and monitor your Python application in production in a simple and intuitive way, but without the need for DevOps skills.Demetrios BrinkmanFounder at MLOps.commmunity
SaaS, Managed & On Premise (local) installation
- Use our secure SaaS service to get started right away with your project
- Prefer to run UbiOps on your own (cloud) infrastructure? UbiOps is also available in an OnPremise version. Suitable for GCP, Azure, AWS or other infrastructure supporting Kubernetes. Installing and managing it on your own clusters is easy
- UbiOps is also available as a managed service on Private Cloud for working with sensitive data. Ask us for the possibilities.