What is the Role of a - Machine Learning Engineer

This course outlines the role's details of a Machine Learning Engineer. You will understand the practical on-the-job details of the role.

ML Engineer

What is a typical day in the life of a Machine Learning Engineer?

1. ML Engineers work more on the engineering and less on the science part of the data problem
2. They work on
• Data Extraction
• Data Cleansing
• Model Building
• Model Deployment

ML Engineer Role

Can you explain with an example?

Say, for example, the business wants to develop an application that can classify the user sentiments on the customer's Twitter handle and create some charts based on the same. An ML Engineer would be doing the following:

• Creating the interface from Customer's Application to Twitter API
• Extracting Data from Twitter API
• Cleaning the Data
• Building a Sentiment Classification Model
• Deploying the Model to Customer's Ecosystem
• Maintaining the application

Big Data & DevOps

Should an ML Engineer have any knowledge of Big Data and DevOps?

Yes, most of the big enterprises maintain and archive their data in a Big Data ecosystem. Knowing how to pull the data and work on top of it will help an ML Engineer.

Do ML Engineers need to know DevOps concepts? If yes, to what extent should they be familiar?

Yes, Absolutely. DevOps has become the de-facto term nowadays. ML Engineers should be able to

• Create CI / CD pipelines
• Write Automation Test Scenarios for the application they have developed
• Monitor the application developed
• Containerize the application

ML Engineer vs Data Scientist

Where do we draw a line between ML Engineer and Data Scientist?

Data scientists work more on the Math and Stats part to develop and fine-tune the algorithms required for a given business scenario. ML Engineers take the model and bring them to life by productionizing them.

Technology

What would be an ideal technology that an ML Engineer should be proficient in?

There is no ideal technology but the developer community is embracing Python. Depending on the customer's needs, one has to familiarize a specific technology. Conceptually, developing the application and deploying it into an ecosystem is similar for many technologies.

Technical Expectations

What am I expected to Learn and Know to become a Machine Learning Engineer?

From a Technology standpoint, you should know the following Libraries in Python to perform ML activities.

• Numpy and Pandas - To perform Data Cleansing and Exploratory Analysis
• Matplotlib - To perform Visualizations
• NLTK - To perform NLP
• scikit learn - For Shallow Learning Algorithms
• flask - To develop APIs

In this journey, you are going to explore all these areas and build upon your skills to become a good Machine Learning Engineer.