Create production ready features without leakage

With Kaskada, you can choose to calculate feature values at any point-in-time to train models without the risk of leakage. When you’re ready, you can compute the same feature values with a time of “now” to make new predictions using a live model in production.

See Our Solutions

Kaskada is the best in the world in time travel, now you can create and operate predictive models with event-based data.

&& (

)

Why Choose Kaskada

Join your data without leakage

Connect to all your event-based data sources

  • Group and re-group data to associate with Entities
  • Create new tables and views as needed
  • Perform temporally accurate joins across Entities

Iterate in your favorite environment

Love Jupyter for experiments? Use our Python client library

  • Leverage notebooks for experimentation
  • Build up features and time selection during exploration
  • Change time selection to affect all features without needing to make data pipeline changes

Share your features as code not just data

Enable sharing, preserve privacy, prevent data leakage

  • Share feature definitions, not just results
  • Easily compute the same features over different datasets
  • Deploy multiple instances of Kaskada's compute engine to support federated learning

Make feature values available for production

Hate Jupyter for production? Use our Python client library

  • Write batch jobs to cache your feature values
  • Integrates with your model orchestration platforms
  • Support streaming when you’re ready

Join the Community!

Get Updates From Kaskada