With Kaskada, you can choose to calculate all feature values at any point in time you choose. Or you can calculate the feature values for each entity at the time an event occurred. For instance, calculate all features at the exact time a user made a large purchase, or at the time of a fraudulent transaction, or when a customer churned out.
Use these point-in-time and event-driven feature values to train time-based models without risk of leakage. When you're ready, you can export features with a time of "now" to make new predictions using a live model in production.
How It Works
Connect to Data
Ongoing access to all the data needed for feature creation
Ingest historical data from data warehouses and data lakes
Explore data and define relationships between event-based data sources
Connect directly to streaming data sources, like Apache Kafka and Amazon Kinesis (Enterprise only)
Design and Visualize Features
Create features using a simple, interactive interface
Write impactful features using precise, point-in-time historical feature values
Visualize feature distributions across time in interactive charts
Commit your work and collaborate with others
Select and Export Features
Discover and share features to train models
Compare feature values visually to find meaningful patterns
Normalize and encode features values with a few clicks
Choose features to export for training, using feature values across time
Share features with your entire team in the feature store.
Deploy to Production
Deliver features directly to customer-facing applications
Set up recurring batch calculations of all feature vectors to use in production
Update the feature versions used in production in minutes
Call the feature store API to receive real-time feature values
Kaskada offers three plans for individuals, data science teams, and organizations to streamline feature engineering. Choose to learn and evaluate using public data with the Community edition, to collaborate and increase your impact using your own data with Data Science edition, and to power dedicated, production-ready infrastructure with real-time data in the Enterprise edition.
Choose the edition that works best for you. For Community Edition and Data Science Edition, fill out the signup form and read and agree to the terms of service.Your application will be reviewed and approved within 2 business days. For Enterprise edition, fill out the contact form and we will reach out to set up a call with you.
What happens after I sign up?
Once your application is approved, you will need to schedule an onboarding session with Max Boyd, Kaskada’s Data Science Lead. Once your onboarding session is complete, you’ll receive an email to set your account password and can then log into Kaskada and begin to use the product!
What if I want to cancel?
If you're on a paid tier of Kaskada, you can cancel at anytime.Just send us a note in the chat widget or reach out via email, and our team can help.There is no charge to be part of the Community or the free tier of the Data Science Edition.Enterprise edition customers should reach out to their Account Manager to inquire about your contract subscription period.