navio

A managed AI platform to train, manage, monitor, and deploy Machine Learning models.

navio is a platform for managing and performing various machine learning operations across a organization's entire AI landscape. From training statistical models from the data in your spreadsheets to speeding up productionization of your custom models by effortlessly spinning up deployments and providing REST endpoints for prediction.

Automated Machine Learning

Data scientists turn raw data into valuable insights by utilizing their analytical and statistical skills to determine how a dataset should be cleaned and which statistical methods or strategies should be leveraged to get the best results when training machine learning models. navio's automatic machine learning capabilities allow users to identify these useful patterns in datasets without prior knowledge of these techniques. Simply upload a dataset, configure the model to train and navio will automatically train a model using SparkML. Once a model has been trained it can be evaluated and instantly deployed for quick integration.

  • available for time-series and
    tabular datasets

  • train binary, multi-class classification, or regression models

  • K-fold validation

  • upload data as CSV or connect to a database via JDBC

  • evaluate model performance

Custom models

Upload any custom python-compatible model to navio to make use of its user interface to easily manage, deploy and integrate your models. navio makes use of MLflow, a standard for packaging ML models so that you can continue to create models using your favorite tools and deep learning frameworks.

MLOps

Whether training your model on navio or uploading your own custom model, navio provides various Machine Learning operations to support you during the model development process all the way to running your model in production. Automatically create REST endpoints and keep track of the machines or clients that are interacting with your model.

One-click deployment

Deploy a model with a single click.

Integration

Use REST endpoints to integrate model predictions in your application

Monitoring

Monitor any interactions with the model.

Retraining

Retrain a model when more data becomes available.

More than just deploying a model

Organization and Simplicity

Stay organized by managing models in various workspaces and use cases. Control access via user management.

Out-of-distribution detection

Out-of-distribution detection (OOD) for detecting outliers in the data during prediction.

Explanations

SHAP explanations to get explainable predictions or configure your own explanation visualizations for your custom models using Plotly.

Machine overview

Have an overview of all machines connected to your deployed models.

10101101010101010

1000110111011

navio API

The navio API allows data engineers and data scientists to programmatically include navio in their development or production workflow. Use the API to perform various tasks including training, uploading, and deploying.

 

Are you interested? Talk to us about navio!