flintml
FlintML is an all-in-one, self-hosted machine learning platform designed for real-world teams seeking a simple, flexible, and fast-to-deploy solution. It centralizes the entire MLOps lifecycle to reduce infrastructure overhead while providing a developer-centric experience. The platform features a Delta Lake storage layer that adds ACID guarantees and time travel to data pipelines, a unified data catalog treating tables, models, and artifacts as first-class citizens, and efficient data processing powered by Polars for memory-efficient operations. Integrated tools include experiment tracking via Aim for comparing runs, a Jupyter Lab environment for seamless notebook development, and workflow orchestration through Dagster. Built on Docker Compose, FlintML offers reproducible compute with support for flexible drivers and composable, declarative infrastructure. It requires Docker and runs on Linux or WSL systems, currently excluding ARM architecture. Users can quickly deploy the platform with a single command to