Quarterly Product Update: Fall

Chalk’s quarterly updates focus on integrating AI agents into engineering workflows while maintaining human oversight and controlled access to data, models, and tools. The Chalk MCP Server and dashboard-based Chalk Assistant allow agents to query sources, run functions and notebook cells, inspect logs and traces, and document investigations in notebooks containing Python, SQL, and Markdown. Chalk Model Training enables teams to train and deploy models using production features, with managed CPU or GPU execution, versioned inputs, checkpoints, and metrics; Grid reportedly used it to reduce model development time by more than half and increase weekly training runs from about 10 to 30. Chalk also introduced Model Gateway for centralized controls such as budgets, rate limits, fallbacks, and model-call tracing, alongside MCP Gateway for policy-governed tool access and auditing without exposing backend credentials. Additional SQL, DataFrame, IDE, Iceberg catalog, and linting capabilities provide consistent interfaces for engineers and agents to explore federated data, validate queries, and move from experimentation to production.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.