Glad you liked it. Optimization is possible absolutely. Airflow helps you create DAG based workflows. It is more on the lines of creating a pipeline and monitoring processes in that pipeline rather than solely monitoring model training parameters as we do in mlflow. Both these frameworks can be combined but. If multiple ML models are applied on a model in sequence, that pipeline can be designed using Airflow and model training and inference parameters can be logged and tracked using MLflow

Engineering at NAYAN Technologies. On a quest for technology.