Modern Data Stack
June 28, 2022
On the heels of announcing our $14.5M Series A and General Availability, we’re excited to be at the Data + AI Summit to unveil support for Continual on the Databricks Lakehouse. Increasingly, data and ML tool providers are embracing a data-centric approach to the ML workflow. The goal is to focus on what increasing drives ML – the data – compared to infrastructure, algorithms, or pipelines.
At Continual we bet on data-centric AI from day one. It's the first MLOps solution to offer a data-first declarative workflow and end-to-end automation of the entire ML lifecycle. You can focus on the data and business objective and Continual helps you automate the rest.
Continual sits directly on top of Databricks and other cloud data platforms, including Redshift, BigQuery, and Snowflake. It lets any data professional build continually-improving predictive models — from customer churn to inventory forecasts — without complex engineering. All your features and predictions stay in your lakehouse, ensuring consistent governance and unified access.
Continual’s secret sauce is in the separation it creates between your use cases, which are written declaratively, and its operational layer.
Continual’s declarative layer includes a feature store for defining and sharing features using SQL and a model store for defining prediction tasks and monitoring models.
Continual’s declarative design means you can seamlessly transition from Continual’s UI for exploration to CLI for operationalization. If you already have dbt models, you can just annotate them with additional metadata and get continually improving predictions in your lakehouse. Continual automatically builds and maintains models for you, letting anyone design and deliver multiple ML solutions using the same GitOps-friendly development pipeline and workflow.
This is just the beginning. We’re excited by the possibilities enabled by Serverless SQL, streaming support, and bi-directional interoperability with MLFlow 2.0. Stay tuned for more announcements later this year.
What's the secret to building a great data team and enabling AI use cases? We'll dive in during this article.