In the first part of the course, we have introduced the importance of having “good” data as well as several machine learning techniques.
Is this enough to create ML applications?
In real life, we have to answer many other questions to take machine learning models to production. Namely:
- What are the business/performance needs? Are there any ethical/legal problems?
- Where will the data come from? In what form is the data? What is the quality of the data? Can we use the data for training?
- Do we need to transform the data? What type of model is required?
- How good is the trained model?
- How will the model be deployed?
- Does the model need to be monitored? If yes, how?
MLOps
Quality of Model
Multiple iterations are often required
Previous Section
Home
Next Section