Snowflake releases Cortex Code for AI/ML pipelines
Snowflake has released Cortex Code (CoCo), a framework for deploying and orchestrating Python-based AI and machine learning pipelines directly within its Data Cloud. The platform is designed to operationalize models and workflows without moving data, supporting both batch and real-time use cases with an emphasis on low-latency deployment and enterprise security.
- Cortex Code is part of Snowflake's broader Cortex AI suite and is available in two forms: a command-line interface (CLI) which is now generally available, and an integration within the Snowsight web interface, which is expected to be generally available soon. - The tool is designed to be "Snowflake-aware," meaning it understands a user's specific data catalog, schemas, and role-based access controls (RBAC), allowing it to generate contextually relevant SQL, dbt, and Python code while respecting existing data governance policies. - For data engineers, specific use cases include generating dbt models with tests, discovering tables with PII for compliance, and optimizing query performance by refactoring SQL, all through natural language prompts. - Unlike generic AI coding assistants, Cortex Code can interact directly with a user's local development environment, including accessing local files to manage dbt projects or Streamlit applications, and can orchestrate local git and bash commands. - The CLI version is billed on a pay-as-you-go basis according to token consumption, while the Snowsight version is currently free of charge, with pricing details to be announced later. - This release intensifies Snowflake's competition with Databricks, which targets similar AI/ML workloads with its Mosaic AI platform; the choice often depends on whether a team is analytics-led (favoring Snowflake's SQL-centric ecosystem) or engineering-led (favoring Databricks' flexibility for complex model development). - Early adopters in the insurance industry, such as Shelter Mutual Insurance, have highlighted the tool's utility in regulated environments, where it helps reduce development friction without disrupting established workflows and governance controls. - Snowflake also announced an upcoming integration with Vercel's v0, which will allow developers to use natural language to build AI-powered data applications that can be deployed securely inside Snowflake using Snowpark Container Services.
