platoseedplatoseed
Open source version control for metadata
Grai is open source version control for metadata. We can determine how database changes will affect deployed machine learning models, apis, and dashboards because we understand how data relates across systems which don’t otherwise talk to each other.
Ian is a recovering ex-physicist turned machine learning engineer with experience building MLE teams everywhere from startups (Bellhops) to fortune 50 companies (Centene). An active open source contributor he's written software with over 11M downloads (Visions). Grai was born from the pain of working on distributed data particularly within enterprise settings.
Grai knows how a database migration will impact deployed ML models, APIs, and more because it understands how data relates across systems.
Grai opens-source data management platform that captures automated data lineage and integrates with git to run data-integrity tests for downstream users as part of CI/CD, with a self-hosted option and a cloud option coming soon. It targets developers and data teams to make testing data flows across the stack easier by centralizing metadata and enabling end-to-end visibility.
From the original launch (Sep 2022) — may be outdated.
Open-source monitoring for machine learning models
Metadata for GenAI workflows
