Unified MLOps: Feature Stores & Model Deployment
Wednesday, January 20 | 
11:40 AM - 
12:10 PM PST

If you’ve brought two or more ML models into production, you know the struggle that comes from the complex process. This talk will teach you a whole new approach to MLOps that allows you to successfully scale your models without increasing latency, by merging a database, a feature store, and machine learning.

Splice Machine is an open-source platform that allows for deployment of machine learning models as intelligent tables inside of our unique hybrid (HTAP) database. When new data is inserted into these tables predictions are automatically generated by, and stored in, the same database table. This integrated structure takes away the need for endpoints and containers, meaning models can be deployed with just one line of code.

The same HTAP database that makes database model deployment possible powers a world class feature store as well. Most feature stores simply tape together an analytical SQL engine and an operational key-value store, resulting in latency and unnecessary data duplication. Splice Machine’s unified SQL architecture allows for better performance, and is easier to use.

In this talk, Monte will discuss how his experience running the AI lab at NASA, and as CEO of Red Pepper, Blue Martini Software and Rocket Fuel, led him to start Splice Machine, and how it solves the MLOps problems many in the industry face today.

Speakers

Jack Ploshnick
Standard
Customer Data Scientist
Splice Machine
Monte Zweben
CEO
Splice Machine
Scroll to Top