With Kaskada, you can connect directly to your event-based data and calculate aggregated feature values at any point in time to train models without the risk of leakage.
When you’re ready, you can compute the current value of the same features to make predictions using the latest data and a live model in production.
- Connect your event-based data
- Build and iterate with familiar interfaces
- Share features as code
- Deploy feature vectors to production
Features are composed using Fenl, a query language designed for authoring and sharing feature definitions.
Fenl expressions are temporal -- they describe how the result of a computation changes over time rather than just the current result. Temporal queries make it easy to reconstruct the information available at arbitrary times in the past.
The rich time-traveling tools provided by Fenl make it easy to build training datasets without leaking knowledge of the future.
Collaborate by sharing code, not just data. Fenl's composability allows you to share and re-use cleaned or pre-aggregated expressions.
Authoring in Fenl ensures your features are scalable, reproducible, and can be computed incrementally.
Features can be exported to external storage for training. When you're ready to host a model, the same feature definitions used in training can be continuously written to a feature-vector cache.
Updated about 2 months ago