Using Airflow with Tensorflow and MLFlow

Watch On Demand

Missed the Webinar? Watch Now.

By proceeding you agree to our Privacy Policy , our Website Terms and to receive emails from Astronomer.


In a previous webinar, we looked at using Apache Airflow for productionizing ML pipelines and showed how Airflow is suitable for orchestrating operationalized ML workflows. Many MLOps tools can be seamlessly integrated with Airflow.

Now, we’ll delve further into how Airflow can be integrated with Tensorflow and MLFlow specifically to manage ML pipelines in production, using a worked example to demonstrate.

Specifically, we’ll showcase the following actions using Airflow:

  • Preparing images for deep learning-based model training.
  • Training an image classifier model using Tensorflow and Ray.
  • Tracking the model training runs using MLFlow.
  • Serving the model as a realtime API end point using Ray.

Hosted By

  • Kenten Danas Kenten Danas Lead Developer Advocate
  • Jeff Fletcher Jeff Fletcher Director of Field Engineering EMEA