Turn any python prediction function into a REST API, deployed on the cloud. API documentation and client libraries are inferred from type annotations and docstrings and automatically published to the web.
No more dev-ops. Write a predict function with type annotations and a docstring. Specify the dependencies (requirements.txt) and which function implements the API (Procfile). That's it.
Get Model-ops best-practice with one command. Run dataka model deploy. to deploy your function as an API endpoint to the cloud. Get Open API documentation, and model versioning for free.
| Name | Type | Description |
|---|---|---|
| crime | float | crime rate per 100K |
| age | int | home age in years |
| water | bool | view of water |
200: model.Prediction400: model.InvalidParamsReduce development time and be a hero for the engineering team. Codegen automatically writes client libraries for your API in your target language and uploads them to the cloud as git repositories. Just install using your native package manager.
Install custom libraries (requirements.txt) or redefine your environment (Dockerfile). Just include the file and we’ll do the rest.
Support for GraphQL and grpc coming soon.
There are many tools claiming to simplify dev-ops for data scientists. Most have steep learning curves and are too complex for data science use cases. Our tools are built by data scientists for data scientists. They are powerful but intuitive to learn.
Published notebooks rerun in the cloud with stable, versioned links to output. All your business logic and data lineage are documented in one single source of truth.
Dataka Office
415 Mission St
26th Floor
San Francisco, CA 94105
Copyright © 2021
All rights reserved.