Introducing PostgresML
PostgresML (https://postgresml.org/) is an extension for Postgres that allows practitioners to implement the entire ML life cycle on top of a Postgres database for text and tabular data.
PostgresML utilizes SQL as the interface to train models, create deployments, and make predictions. The use of SQL means model and data operations can be combined seamlessly and fit naturally into Postgres DB data engineering environments.
There are many advantages to having a shared data and ML platform. As we saw in the previous chapter, with SageMaker, significant effort is spent on moving data around. This is a common problem in ML environments where data, especially transactional data, lives in production databases, and complex data engineering workflows need to be created to extract data from production sources, transform the data for ML use, and load the data into a store that’s accessible to the ML platform (such as S3 for SageMaker).
By combining the...
