Today, the team at Dagster Labs is proud to announce an early release of Dagster, an open-source library for building systems like ETL processes and ML pipelines. Dagster defines data applications as graphs of functional computations that produce and consume data assets. These applications are queryable and operable via an API, allowing builders to use any tool they choose while collaborating on the same logical application. The computational graphs are abstract and can be deployed to arbitrary compute targets, including Airflow, Dask, Kubernetes-based workflow engines, and FaaS platforms. Adopting Dagster is expected to improve productivity, testability, reliability, and collaboration in data applications, leading to an entirely new open ecosystem of reusable data components and shared tooling. The library aims to address the pain points of building modern data applications, including uncontrollable inputs, multi-persona and multi-tool nature, and difficulty in testing and developing these systems. Dagster's core model is inspired by functional programming principles and provides a coarse-grained abstraction for data computations. It offers features such as queryable, operable, and monitor-able data applications, self-describing metadata, business logic defined in the user's tool of choice, and executable in various environments. The team behind Dagster is looking for partners to collaborate on building complex data applications and is also hiring for its growing team.