If you are a data scientist, you definitely have felt the need to keep track of your machine learning models. If you are not careful, you can quickly lose track of the best model you have been tuning for hours.

Numerous platforms can help you deal with this issue. In this post, we will focus on the open-source project called MLflow and how we implemented it to make our model training efforts more efficient.

According to MLflow’s documentation:

MLflow is an open source platform for managing the end-to-end machine learning lifecycle.

The platform facilitates workflow management for training, tracking, and producing machine learning models.

This post will not focus on what is MLflow and why you should use it because countless websites already explain that. Here, we will implement it on AWS using technologies, such as Docker, ECR, and EC2. We assume that you are already familiar with them.

In the end, you will access and use the MLflow UI hosted on a remote server.

You can check the full blog post on my Medium: Set Up MLflow on AWS EC2 Using Docker, S3, and RDS