Worked on the ecmwf/anemoi-core repository to enhance ML experiment tracking by improving the MLflow schema and logger validation. Developed a schema extension in Python using Pydantic, adding log_hyperparams and prefix fields to align configuration validation with runtime behavior. Implemented comprehensive unit tests to ensure that constructor parameters for AnemoiMLflowLogger and AnemoiAzureMLflowLogger remain consistent with the schema, reducing the risk of future regressions. Addressed a configuration validation issue that previously rejected valid hyperparameter logging inputs, thereby improving reliability and cross-platform compatibility for MLflow and Azure MLflow integrations. Focused on backend development and robust testing practices throughout the process.