Egor Stopalov.
ML systems, like any software product, have a life cycle: a task is chosen, data is collected, one or more models are trained. Then the system needs to be made available to users. By inference, the author means the process in which the system receives data from the user, makes predictions, and responds. He also uses the same word for the code and infrastructure that support this process. The talk shows three generations of such infrastructure in a medical ML startup.
Continue reading — it's free
Google or a one-time email code. No password.