Roles in MLOps

1. Roles in MLOps

Well done on the course so far! Let's now explore the different roles in the machine learning lifecycle.

2. Machine learning lifecycle

From the phases we've seen, we can outline a sequence of steps that the lifecycle goes through. Each task requires different roles, which can be categorized as business roles and technical roles.

3. Business roles

In the business category, there are two main roles: business stakeholder and subject matter expert.

4. Business roles: business stakeholder

The business stakeholder, or product owner, is a managerial staff member making budget decisions and ensuring the machine learning project aligns with the company's vision. They are involved throughout the lifecycle, defining business requirements during the design phase, assessing initial experiment results in the development phase, and verifying outcomes in the deployment phase.

5. Business roles: subject matter expert

Secondly, there is the subject matter expert, who has domain knowledge about the problem that we are trying to solve. The subject matter expert is involved throughout the lifecycle because they can assist the more technical roles with interpreting the data and results at each step.

6. Technical roles

The main technical roles are data scientist, data engineer, and ML engineer.

7. Technical roles: data scientist

The data scientist is responsible for data analysis and model training and evaluation. The evaluation includes monitoring the model once it has been deployed to ensure that the model predictions are valid. We can find the data scientist in all phases of the lifecycle, but mostly during the development phase.

8. Technical roles: data engineer

The data engineer is responsible for the collecting, storing, and processing of data. This also means that the data engineer should check the data quality and include tests such that the quality is maintained throughout the process. Therefore, the data engineer is mostly involved with tasks that have to do with data before training the model, during the model training, and once the model is used in production.

9. Technical roles: ML engineer

The machine learning engineer is a relatively new role that is quite versatile and designed specifically to have expertise over the entire machine learning lifecycle. It is a cross-functional role that overlaps with the other technical roles. As such, the machine learning engineer is involved in all phases. They know, for instance, how to extract and store data and develop or deploy a machine learning model.

10. Additional roles involved in ML

Beyond the primary roles mentioned earlier, we can often find various other roles involved in the machine learning lifecycle. Roles such as data analysts, developers, software engineers, and backend engineers can all contribute significantly depending on the specific application of machine learning and the type of company. For instance, in a startup, roles may be more versatile, whereas in an enterprise, roles tend to be more defined and specialized.

11. Let's practice!

Let's dive into some exercises to practice with the roles we have just looked into!

Create Your Free Account

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.