What is expected from ML engineers?

ML engineering involves using ML and software engineering concepts and techniques to design, build, and manage production-level ML systems, along with pipelines. In a team working to build ML-powered applications, ML engineers are generally expected to build and operate the ML infrastructure that’s used to train and deploy models. In some cases, data scientists may also need to work on infrastructure-related requirements, especially if there is no clear delineation between the roles and responsibilities of ML engineers and data scientists in an organization.

There are several things an ML engineer should consider when designing and building ML systems and platforms. These would include the quality of the deployed ML model, along with the security, scalability, evolvability, stability, and overall cost of the ML infrastructure used. In this book, we will discuss the different strategies and best practices to achieve the different objectives of an ML engineer.

ML engineers should also be capable of designing and building automated ML workflows using a variety of solutions. Deployed models degrade over time and model retraining becomes essential in ensuring the quality of deployed ML models. Having automated ML pipelines in place helps enable automated model retraining and deployment.

Important note

If you are excited to learn more about how to build custom ML pipelines on AWS, then you should check out the last section of this book: Designing and building end-to-end MLOps pipelines. You should find several chapters dedicated to deploying complex ML pipelines on AWS!