Search papers, labs, and topics across Lattice.
GitLake introduces a Git-inspired framework tailored for data management in agent-first lakehouses, enabling the creation of commits, branches, and merges across multiple tables. This system allows agents to operate on isolated branches while facilitating human oversight during the review and publication process, ensuring atomic visibility of outputs. Key findings include production lessons and correctness insights derived from a preliminary Alloy model, highlighting the system's robustness and potential for enhancing collaborative data workflows.
Agents can now work independently on data changes while humans maintain oversight, revolutionizing collaborative data management.
We present GitLake, a Git-for-data design for an agent-first lakehouse. The system lifts single-table Iceberg snapshots into lakehouse-wide commits, branches, and merges, letting agents work on isolated branches while humans review and publish changes. Pipelines run on temporary branches and publish through a final merge, so all outputs become visible atomically or none do. Finally, we report production lessons as well as correctness insights from a preliminary Alloy model of our core abstractions.
CMU ML