Systems for processing data using stages

Systems and techniques for processing data using data pipelines. For instance, a first version of data may be read from one or more databases for processing by a data pipeline. Stages of the data pipeline may then process the data, such as by transforming, validating, or changing the data, in order to generate additional versions of the data. These versions of the data are passed through the subsequent stages of the data pipeline. Additionally, metadata may be generated for one or more stages of the data pipeline, where the metadata represents information describing the processing by the one or more stages. This metadata may also be passed through the subsequent stages of the data pipeline. Based on the last stage of the data pipeline processing the data, the versions of the data along with the metadata may then be written into the one or more databases.