Lightning AI Studios: Never set up a local environment again →

Log in or create a free account to track your progress and access additional course materials  

10.3 Designing Machine Learning Systems

In this lecture, we learned about the comprehensive process of designing machine learning systems, starting with understanding the requirements. We referenced the importance of outlining business, user, and performance requirements to guide the structure of the system. We also discussed how to formulate machine learning problems based on these requirements, determining whether machine learning is necessary, defining the model’s inputs and outputs, and choosing the type of model to use.

Lastly, we discussed the need to monitor and maintain the deployed model, ensuring it remains accurate and can be improved upon, under considerations like data drift, increasing accuracy, or reducing latency.

Additional resources if you want to learn more

Unfortunately, we can’t cover these topics in more depth as each of them would be a course in itself. However, if you are interested in additional references, I can recommend Chip Huyen’s book on Designing Machine Learnign Systems, for example.

Log in or create a free account to access:

  • Quizzes
  • Completion badges
  • Progress tracking
  • Additional downloadable content
  • Additional AI education resources
  • Notifications when new units are released
  • Free cloud computing credits

Quiz: 10.3 Designing Machine Learning Systems

What is the first step in defining a machine learning problem?

Incorrect. This would be a good 2nd step though.

Incorrect. Although, this becomes a core component later.

Correct. It’s always recommended to define the problem to be solved (and evaluate whether machine learning is even needed.)

Incorrect. This is one of the last steps (before deployment).

Please answer all questions to proceed.
Watch Video 1

Unit 10.3